devlyn-cli 2.2.2 → 2.3.1

package/scripts/lint-skills.sh

@@ -18,6 +18,113 @@ section() { printf '\n%s=== %s ===%s\n' "$dim" "$1" "$reset"; }
 ok()      { printf '  %s✓%s %s\n' "$green" "$reset" "$1"; }
 bad()     { printf '  %s✗%s %s\n' "$red"   "$reset" "$1"; fail=1; }
 
+make_temp_file() {
+  local __var="$1"
+  shift || true
+  local path
+  if ! path=$(command mktemp "$@"); then
+    bad "mktemp failed: ${*:-<default>}"
+    return 1
+  fi
+  printf -v "$__var" '%s' "$path"
+}
+
+make_temp_dir() {
+  local __var="$1"
+  shift || true
+  local path
+  if ! path=$(command mktemp -d "$@"); then
+    bad "mktemp -d failed: ${*:-<default>}"
+    return 1
+  fi
+  printf -v "$__var" '%s' "$path"
+}
+
+section "Check 0a: Temp allocation fails closed"
+direct_mktemp=$(grep -nE '(^|[ =])mktemp( |$)|\$\([[:space:]]*mktemp' scripts/lint-skills.sh \
+  | grep -v 'command mktemp' \
+  | grep -v 'make_temp_' \
+  | grep -v 'direct_mktemp=' || true)
+if [ -z "$direct_mktemp" ]; then
+  ok "lint-skills.sh uses temp allocation helpers instead of direct mktemp"
+else
+  while IFS= read -r f; do bad "$f"; done <<< "$direct_mktemp"
+fi
+
+# iter-0034 Phase 4 cutover (2026-05-03): legacy skill paths dropped.
+# Surface is the 2-skill product (`/devlyn:resolve` + `/devlyn:ideate`)
+# plus the `_shared/` kernel. Keep this list single-source so all installed
+# mirror parity checks cover the same files.
+critical_path_files=$(cat <<'EOF'
+_shared/spec-verify-check.py
+_shared/collect-codex-findings.py
+_shared/verify-merge-findings.py
+devlyn:ideate/SKILL.md
+devlyn:ideate/references/spec-template.md
+devlyn:ideate/references/elicitation.md
+devlyn:ideate/references/project-mode.md
+devlyn:ideate/references/from-spec-mode.md
+devlyn:resolve/SKILL.md
+devlyn:resolve/references/state-schema.md
+devlyn:resolve/references/free-form-mode.md
+devlyn:resolve/references/phases/plan.md
+devlyn:resolve/references/phases/probe-derive.md
+devlyn:resolve/references/phases/implement.md
+devlyn:resolve/references/phases/build-gate.md
+devlyn:resolve/references/phases/cleanup.md
+devlyn:resolve/references/phases/verify.md
+_shared/expected.schema.json
+_shared/adapters/README.md
+_shared/adapters/opus-4-7.md
+_shared/adapters/gpt-5-5.md
+_shared/codex-config.md
+_shared/codex-monitored.sh
+_shared/engine-preflight.md
+_shared/pair-plan-schema.md
+_shared/runtime-principles.md
+EOF
+)
+
+check_skill_mirror_parity() {
+  local target_dir="$1"
+  local skip_msg="$2"
+  local missing_prefix="$3"
+  local differ_suffix="$4"
+  local ok_msg="$5"
+  local drift=0
+  local rel src dst
+
+  if [ ! -d "$target_dir" ]; then
+    ok "$skip_msg"
+    return
+  fi
+
+  while IFS= read -r rel; do
+    [ -n "$rel" ] || continue
+    src="config/skills/$rel"
+    dst="$target_dir/$rel"
+    if [ ! -f "$src" ] || [ ! -f "$dst" ]; then
+      bad "$missing_prefix: $rel"; drift=1; continue
+    fi
+    if ! diff -q "$src" "$dst" >/dev/null 2>&1; then
+      bad "$rel — $differ_suffix"
+      drift=1
+    fi
+  done <<< "$critical_path_files"
+
+  # iter-0009: codex-monitored.sh must be executable in installed mirrors
+  # (skills trees get copied into work dirs for variant arms; bash refuses to
+  # run a non-executable wrapper).
+  if [ -f "$target_dir/_shared/codex-monitored.sh" ] \
+     && [ ! -x "$target_dir/_shared/codex-monitored.sh" ]; then
+    bad "_shared/codex-monitored.sh — not executable in ${target_dir} mirror"
+    drift=1
+  fi
+  if [ $drift -eq 0 ]; then
+    ok "$ok_msg"
+  fi
+}
+
 # ---------------------------------------------------------------------------
 # 1. No MCP references in managed source or user-facing docs.
 # ---------------------------------------------------------------------------
@@ -69,6 +176,18 @@ else
   while IFS= read -r f; do bad "$f"; done <<< "$offenders"
 fi
 
+# ---------------------------------------------------------------------------
+# 2a. Packaged root instruction files must not contain pyx-memory secrets.
+# ---------------------------------------------------------------------------
+section "Check 2a: No pyx-memory secrets in packaged root instructions"
+offenders=$(grep -RInE 'memory\.api\.pyxmate\.com|Authorization: Bearer pyx_[A-Za-z0-9]{12,}|X-API-Key: pyx_[A-Za-z0-9]{12,}|pyx_[A-Za-z0-9]{16,}' \
+  AGENTS.md CLAUDE.md 2>/dev/null || true)
+if [ -z "$offenders" ]; then
+  ok "AGENTS.md and CLAUDE.md contain no pyx-memory secret material"
+else
+  while IFS= read -r f; do bad "$f"; done <<< "$offenders"
+fi
+
 # ---------------------------------------------------------------------------
 # 3. No stale model strings (gpt-5.0..5.4 hardcoded outside config).
 # ---------------------------------------------------------------------------
@@ -121,66 +240,54 @@ if [ $missing -eq 0 ]; then
   ok "all devlyn:* skills have name: field"
 fi
 
+# ---------------------------------------------------------------------------
+# 5a. devlyn:design-ui is a required skill, not an optional addon.
+# ---------------------------------------------------------------------------
+section "Check 5a: devlyn:design-ui is required"
+if [ -f "config/skills/devlyn:design-ui/SKILL.md" ]; then
+  ok "devlyn:design-ui source lives in config/skills"
+else
+  bad "devlyn:design-ui must be a required skill under config/skills"
+fi
+if [ ! -e "optional-skills/devlyn:design-ui" ]; then
+  ok "devlyn:design-ui is not in optional-skills"
+else
+  bad "devlyn:design-ui must not be installed as an optional addon"
+fi
+if grep -Fq "skillsToInstall: ['devlyn:resolve', 'devlyn:ideate', 'devlyn:design-ui', '_shared']" bin/devlyn.js; then
+  ok "Codex install includes devlyn:design-ui"
+else
+  bad "Codex skillsToInstall must include devlyn:design-ui"
+fi
+if ! grep -F "name: 'devlyn:design-ui'" bin/devlyn.js >/dev/null 2>&1; then
+  ok "devlyn:design-ui is absent from OPTIONAL_ADDONS"
+else
+  bad "devlyn:design-ui must not be listed in OPTIONAL_ADDONS"
+fi
+
 # ---------------------------------------------------------------------------
 # 6. Source ↔ installed mirror parity on critical path.
 # Only runs if .claude/skills exists (i.e. installer has been run).
 # ---------------------------------------------------------------------------
 section "Check 6: Source ↔ installed mirror parity (critical path)"
-if [ ! -d .claude/skills ]; then
-  ok "no .claude/skills (fresh checkout) — skipping parity check"
-else
-  drift=0
-  # iter-0034 Phase 4 cutover (2026-05-03): legacy skill paths dropped.
-  # Surface is the 2-skill product (`/devlyn:resolve` + `/devlyn:ideate`)
-  # plus the `_shared/` kernel.
-  for rel in \
-      _shared/spec-verify-check.py \
-      _shared/collect-codex-findings.py \
-      _shared/verify-merge-findings.py \
-      devlyn:ideate/SKILL.md \
-      devlyn:ideate/references/spec-template.md \
-      devlyn:ideate/references/elicitation.md \
-      devlyn:ideate/references/project-mode.md \
-      devlyn:ideate/references/from-spec-mode.md \
-      devlyn:resolve/SKILL.md \
-      devlyn:resolve/references/state-schema.md \
-      devlyn:resolve/references/free-form-mode.md \
-      devlyn:resolve/references/phases/plan.md \
-      devlyn:resolve/references/phases/probe-derive.md \
-      devlyn:resolve/references/phases/implement.md \
-      devlyn:resolve/references/phases/build-gate.md \
-      devlyn:resolve/references/phases/cleanup.md \
-      devlyn:resolve/references/phases/verify.md \
-      _shared/expected.schema.json \
-      _shared/adapters/README.md \
-      _shared/adapters/opus-4-7.md \
-      _shared/adapters/gpt-5-5.md \
-      _shared/codex-config.md \
-      _shared/codex-monitored.sh \
-      _shared/pair-plan-schema.md \
-      _shared/runtime-principles.md; do
-    src="config/skills/$rel"
-    dst=".claude/skills/$rel"
-    if [ ! -f "$src" ] || [ ! -f "$dst" ]; then
-      bad "missing file on critical path: $rel"; drift=1; continue
-    fi
-    if ! diff -q "$src" "$dst" >/dev/null 2>&1; then
-      bad "$rel — source and installed differ"
-      drift=1
-    fi
-  done
-  # iter-0009: codex-monitored.sh must be executable in the installed mirror
-  # (skills tree gets cp -R'd into $WORK_DIR for the variant arm; bash will
-  # refuse to run a non-executable wrapper).
-  if [ -f ".claude/skills/_shared/codex-monitored.sh" ] \
-     && [ ! -x ".claude/skills/_shared/codex-monitored.sh" ]; then
-    bad "_shared/codex-monitored.sh — not executable in installed mirror"
-    drift=1
-  fi
-  if [ $drift -eq 0 ]; then
-    ok "critical path parity clean"
-  fi
-fi
+check_skill_mirror_parity \
+  ".claude/skills" \
+  "no .claude/skills (fresh checkout) — skipping parity check" \
+  "missing file on critical path" \
+  "source and installed differ" \
+  "critical path parity clean"
+
+# Codex / agent runtimes in this repo can also expose a project-local
+# `.agents/skills` mirror. If it exists, keep the same critical path in parity;
+# otherwise a session can read stale pair/risk-probe contracts even while the
+# source and `.claude/skills` mirrors are clean.
+section "Check 6a: Source ↔ .agents mirror parity (critical path)"
+check_skill_mirror_parity \
+  ".agents/skills" \
+  "no .agents/skills — skipping parity check" \
+  "missing .agents critical-path file" \
+  "source and .agents mirror differ" \
+  ".agents critical path parity clean"
 
 # ---------------------------------------------------------------------------
 # 6b. VERIFY merge verdict binding self-test.
@@ -194,6 +301,83 @@ if python3 config/skills/_shared/verify-merge-findings.py --self-test >/dev/null
 else
   bad "verify-merge-findings.py self-test failed"
 fi
+if ! grep -Fq 'def pair_trigger_skip_contract_violation' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'def pair_trigger_missing_contract_violation' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'def pair_trigger_reason_completeness_violation' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'def pair_trigger_present' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'KNOWN_PAIR_TRIGGER_REASONS = {' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'mode.pair-verify' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'state.get("pair_verify") is True' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'def pair_flag_contract_violation' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'verify-pair-trigger-conflicting-pair-flags' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq -- '--pair-verify and --no-pair are mutually exclusive' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'def has_known_pair_trigger_reason' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'def all_known_pair_trigger_reasons' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'return reason in KNOWN_PAIR_TRIGGER_REASONS' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq '"reasons": ["risk.high", 3]' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq '"reasons": ["risk.high", "looks-hard"]' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq '"reasons": ["risk high"]' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq '"reasons": ["risk_profile.high_risk", "risk_probes_enabled"]' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'verify-pair-trigger-reasons-unknown' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'verify-pair-trigger-reasons-incomplete' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'pair_trigger.reasons is missing applicable canonical reason(s)' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'pair_trigger.reasons must include a known pair-trigger reason' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'pair_trigger.reasons must only include known pair-trigger reasons' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'verify-pair-trigger-ineligible-unjustified' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'VERIFY state requires a pair decision' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'verify-pair-trigger-user-no-pair-unsupported' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'risk_profile.pair_default_enabled false from an explicit --no-pair opt-out' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'def risk_profile_contract_violation' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'verify-risk-profile-malformed' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'risk_profile.risk_probes_enabled must be a boolean' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'risk_profile.reasons must be a list of strings' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'def spec_frontmatter_complexity' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq '"complexity": "large"' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'complexity.large' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'spec.complexity.high' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'spec.complexity.large' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'verify-pair-trigger-required-missing' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'verify-pair-trigger-skipped-reason-unsupported' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'verify-pair-trigger-mechanical-blocker-unsupported' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'verify-pair-trigger-primary-judge-blocker-unsupported' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'eligible:false` with no supported skip reason' config/skills/devlyn:resolve/references/state-schema.md \
+  || ! grep -Fq 'Canonical eligible reasons are `mode.verify-only`' config/skills/devlyn:resolve/references/state-schema.md \
+  || ! grep -Fq '`mode.pair-verify`' config/skills/devlyn:resolve/references/state-schema.md \
+  || ! grep -Fq 'Eligible triggers must contain only canonical' config/skills/devlyn:resolve/references/phases/verify.md \
+  || ! grep -Fq 'include every applicable canonical reason' config/skills/devlyn:resolve/references/phases/verify.md \
+  || ! grep -Fq '`mode.pair-verify`' config/skills/devlyn:resolve/references/phases/verify.md \
+  || ! grep -Fq '"pair_verify": false' config/skills/devlyn:resolve/references/state-schema.md \
+  || ! grep -Fq 'requires a non-empty reasons list containing every applicable canonical eligible reason' config/skills/devlyn:resolve/references/state-schema.md \
+  || ! grep -Fq 'containing every applicable canonical eligible reason' config/skills/devlyn:resolve/references/state-schema.md \
+  || ! grep -Fq 'state.pair_verify == true' config/skills/devlyn:resolve/references/state-schema.md \
+  || ! grep -Fq 'pair_verify: true` only when `--pair-verify` was passed' config/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'include every applicable canonical reason' config/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq '`--pair-verify` and `--no-pair` are mutually exclusive' config/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'mutually exclusive with `risk_profile.pair_default_enabled == false`' config/skills/devlyn:resolve/references/state-schema.md \
+  || ! grep -Fq 'if both are present, stop with `BLOCKED:invalid-flags`' config/skills/devlyn:resolve/references/phases/verify.md \
+  || ! grep -Fq 'Contradictory, incomplete, or unknown trigger state is a VERIFY contract violation' config/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'user_no_pair` is valid only when `risk_profile.pair_default_enabled == false`' config/skills/devlyn:resolve/references/state-schema.md \
+  || ! grep -Fq 'def reject_json_constant' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'loads_strict_json(raw)' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'invalid JSON numeric constant: NaN' config/skills/_shared/verify-merge-findings.py; then
+  bad "verify-merge-findings.py must block missing pair_trigger and unsupported skip reasons"
+fi
+verify_merge_risk_profile_guard_missing=0
+for file in \
+  config/skills/_shared/verify-merge-findings.py \
+  .claude/skills/_shared/verify-merge-findings.py \
+  .agents/skills/_shared/verify-merge-findings.py; do
+  if ! grep -Fq 'def risk_profile_contract_violation' "$file" \
+    || ! grep -Fq 'verify-risk-profile-malformed' "$file" \
+    || ! grep -Fq 'risk_profile.risk_probes_enabled must be a boolean' "$file" \
+    || ! grep -Fq 'risk_profile.reasons must be a list of strings' "$file"; then
+    bad "$file — verify-merge-findings.py must fail closed on malformed risk_profile"
+    verify_merge_risk_profile_guard_missing=1
+  fi
+done
+if [ $verify_merge_risk_profile_guard_missing -eq 0 ]; then
+  ok "verify-merge-findings.py risk_profile shape guard is mirrored"
+fi
 
 section "Check 6c: Codex stdout collection writes canonical pair findings"
 if python3 config/skills/_shared/collect-codex-findings.py --self-test >/dev/null 2>&1; then
@@ -201,12 +385,1001 @@ if python3 config/skills/_shared/collect-codex-findings.py --self-test >/dev/nul
 else
   bad "collect-codex-findings.py self-test failed"
 fi
+if ! grep -Fq 'def reject_json_constant' config/skills/_shared/collect-codex-findings.py \
+  || ! grep -Fq 'loads_strict_json(raw)' config/skills/_shared/collect-codex-findings.py \
+  || ! grep -Fq 'NaN Codex stdout finding must not normalize' config/skills/_shared/collect-codex-findings.py; then
+  bad "collect-codex-findings.py must reject non-standard JSON constants in pair-JUDGE stdout"
+fi
+
+section "Check 6c1: Archive preserves pair/risk-probe artifacts safely"
+if python3 config/skills/_shared/archive_run.py --self-test >/dev/null 2>&1; then
+  ok "archive_run.py self-test passed"
+else
+  bad "archive_run.py self-test failed"
+fi
+if ! grep -Fq 'SAFE_RUN_ID_RE' config/skills/_shared/archive_run.py \
+  || ! grep -Fq 'run_id must match [A-Za-z0-9_.-]+' config/skills/_shared/archive_run.py \
+  || ! grep -Fq 'Archive devlyn:resolve run artifacts' config/skills/_shared/archive_run.py \
+  || grep -Fq 'Archive auto-resolve run artifacts' config/skills/_shared/archive_run.py \
+  || ! grep -Fq 'invalid JSON numeric constant: NaN' config/skills/_shared/archive_run.py \
+  || ! grep -Fq '"verify.pair.findings.jsonl"' config/skills/_shared/archive_run.py \
+  || ! grep -Fq '"verify-merge.summary.json"' config/skills/_shared/archive_run.py \
+  || ! grep -Fq '"codex-judge.*"' config/skills/_shared/archive_run.py; then
+  bad "archive_run.py must safely archive pair/risk-probe evidence and reject unsafe run ids"
+fi
 
 section "Check 6d: Spec verification executes hidden-blind risk probes"
 if python3 config/skills/_shared/spec-verify-check.py --self-test >/dev/null 2>&1; then
-  ok "spec-verify-check.py risk-probe self-test passed"
+  ok "spec-verify-check.py risk-probe and expected-contract self-test passed"
+else
+  bad "spec-verify-check.py risk-probe / expected-contract self-test failed"
+fi
+if ! grep -Fq 'def validate_present_spec_complexity' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'SPEC_COMPLEXITY_VALUES = {"trivial", "medium", "high", "large"}' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'frontmatter complexity must be one of' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'def validate_sibling_spec_complexity' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'unsupported spec complexity was accepted' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'unsupported sibling spec complexity was accepted by --check-expected' config/skills/_shared/spec-verify-check.py; then
+  bad "spec-verify-check.py checks must reject unsupported spec complexity values"
+else
+  ok "spec-verify-check.py checks reject unsupported spec complexity values"
+fi
+if ! grep -Fq 'generated criteria carrier was not staged into .devlyn/spec-verify.json' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'spec source with mismatched source.spec_sha256 was accepted' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'spec source with matching source.spec_sha256 was not staged' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'source.spec_sha256 mismatch' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'generated criteria without a JSON carrier was accepted' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'generated criteria without source.criteria_sha256 was accepted' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'generated criteria with mismatched source.criteria_sha256 was accepted' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'def source_integrity_error' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'source.criteria_sha256 mismatch' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'Generated criteria were written without one' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'generated criteria carrier was not staged into .devlyn/spec-verify.json' .claude/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'spec source with mismatched source.spec_sha256 was accepted' .claude/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'spec source with matching source.spec_sha256 was not staged' .claude/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'source.spec_sha256 mismatch' .claude/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'generated criteria without a JSON carrier was accepted' .claude/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'generated criteria without source.criteria_sha256 was accepted' .claude/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'generated criteria with mismatched source.criteria_sha256 was accepted' .claude/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'def source_integrity_error' .claude/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'source.criteria_sha256 mismatch' .claude/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'Generated criteria were written without one' .claude/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'generated criteria carrier was not staged into .devlyn/spec-verify.json' .agents/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'spec source with mismatched source.spec_sha256 was accepted' .agents/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'spec source with matching source.spec_sha256 was not staged' .agents/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'source.spec_sha256 mismatch' .agents/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'generated criteria without a JSON carrier was accepted' .agents/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'generated criteria without source.criteria_sha256 was accepted' .agents/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'generated criteria with mismatched source.criteria_sha256 was accepted' .agents/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'def source_integrity_error' .agents/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'source.criteria_sha256 mismatch' .agents/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'Generated criteria were written without one' .agents/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq '"criteria_sha256": generated_hash' .agents/skills/_shared/spec-verify-check.py; then
+  bad "spec-verify-check.py self-test must cover generated criteria source extraction"
+else
+  ok "spec-verify-check.py covers generated criteria source extraction"
+fi
+if ! grep -Fq 'def validate_present_solo_headroom_hypothesis' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'def state_requires_risk_probes' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'def risk_probes_state_error' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq -- '--include-risk-probes accepted missing required risk-probes.jsonl' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq -- '--include-risk-probes accepted non-boolean risk_probes_enabled' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq -- '--include-risk-probes accepted non-object risk_profile' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'def validate_risk_probes_cover_solo_headroom_hypothesis' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'def has_backticked_observable_miss_command' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'backticked command/observable line that exposes the miss' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'weak solo-headroom hypothesis was accepted by --check' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'descriptive backtick solo-headroom hypothesis was accepted by --check' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'risk probe missing solo-headroom command coverage was accepted' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'risk probe with unrelated solo-headroom derived_from was accepted' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'risk-probes[0].derived_from must reference the solo-headroom hypothesis bullet' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'solo-headroom command in a later risk probe was accepted' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'solo-headroom command prefix match was accepted' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq '(?<![A-Za-z0-9_.:/=-])' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'risk-probes[0].cmd must contain a solo-headroom hypothesis observable command' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'weak sibling solo-headroom hypothesis was accepted by --check-expected' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'docs-style solo-headroom hypothesis was rejected by --check' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'docs-style sibling solo-headroom command was rejected by --check-expected' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'def validate_present_solo_headroom_hypothesis' .claude/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'def state_requires_risk_probes' .claude/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'def risk_probes_state_error' .claude/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq -- '--include-risk-probes accepted missing required risk-probes.jsonl' .claude/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq -- '--include-risk-probes accepted non-boolean risk_probes_enabled' .claude/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq -- '--include-risk-probes accepted non-object risk_profile' .claude/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'def validate_risk_probes_cover_solo_headroom_hypothesis' .claude/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'def has_backticked_observable_miss_command' .claude/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'backticked command/observable line that exposes the miss' .claude/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'weak solo-headroom hypothesis was accepted by --check' .claude/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'descriptive backtick solo-headroom hypothesis was accepted by --check' .claude/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'risk probe missing solo-headroom command coverage was accepted' .claude/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'risk probe with unrelated solo-headroom derived_from was accepted' .claude/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'risk-probes[0].derived_from must reference the solo-headroom hypothesis bullet' .claude/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'solo-headroom command in a later risk probe was accepted' .claude/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'solo-headroom command prefix match was accepted' .claude/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq '(?<![A-Za-z0-9_.:/=-])' .claude/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'risk-probes[0].cmd must contain a solo-headroom hypothesis observable command' .claude/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'weak sibling solo-headroom hypothesis was accepted by --check-expected' .claude/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'docs-style solo-headroom hypothesis was rejected by --check' .claude/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'docs-style sibling solo-headroom command was rejected by --check-expected' .claude/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'def validate_present_solo_headroom_hypothesis' .agents/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'def state_requires_risk_probes' .agents/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'def risk_probes_state_error' .agents/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq -- '--include-risk-probes accepted missing required risk-probes.jsonl' .agents/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq -- '--include-risk-probes accepted non-boolean risk_probes_enabled' .agents/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq -- '--include-risk-probes accepted non-object risk_profile' .agents/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'def validate_risk_probes_cover_solo_headroom_hypothesis' .agents/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'def has_backticked_observable_miss_command' .agents/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'backticked command/observable line that exposes the miss' .agents/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'weak solo-headroom hypothesis was accepted by --check' .agents/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'descriptive backtick solo-headroom hypothesis was accepted by --check' .agents/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'risk probe missing solo-headroom command coverage was accepted' .agents/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'risk probe with unrelated solo-headroom derived_from was accepted' .agents/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'risk-probes[0].derived_from must reference the solo-headroom hypothesis bullet' .agents/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'solo-headroom command in a later risk probe was accepted' .agents/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'solo-headroom command prefix match was accepted' .agents/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq '(?<![A-Za-z0-9_.:/=-])' .agents/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'risk-probes[0].cmd must contain a solo-headroom hypothesis observable command' .agents/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'weak sibling solo-headroom hypothesis was accepted by --check-expected' .agents/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'docs-style solo-headroom hypothesis was rejected by --check' .agents/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'docs-style sibling solo-headroom command was rejected by --check-expected' .agents/skills/_shared/spec-verify-check.py; then
+  bad "spec-verify-check.py --check and --check-expected must reject weak solo-headroom hypotheses"
+else
+  ok "spec-verify-check.py rejects weak solo-headroom hypotheses"
+fi
+if ! grep -Fq 'requires `.devlyn/risk-probes.jsonl`' config/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'requires `.devlyn/risk-probes.jsonl`' .claude/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'requires `.devlyn/risk-probes.jsonl`' .agents/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'missing `.devlyn/risk-probes.jsonl` is a CRITICAL mechanical blocker' config/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'missing `.devlyn/risk-probes.jsonl` is a CRITICAL mechanical blocker' .claude/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'missing `.devlyn/risk-probes.jsonl` is a CRITICAL mechanical blocker' .agents/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'requires that file when `state.risk_profile.risk_probes_enabled == true`' config/skills/devlyn:resolve/references/phases/build-gate.md \
+  || ! grep -Fq 'requires that file when `state.risk_profile.risk_probes_enabled == true`' .claude/skills/devlyn:resolve/references/phases/build-gate.md \
+  || ! grep -Fq 'requires that file when `state.risk_profile.risk_probes_enabled == true`' .agents/skills/devlyn:resolve/references/phases/build-gate.md \
+  || ! grep -Fq 'Malformed `state.risk_profile` is also CRITICAL because it can hide enabled risk probes' config/skills/devlyn:resolve/references/phases/build-gate.md \
+  || ! grep -Fq 'Malformed `state.risk_profile` is also CRITICAL because it can hide enabled risk probes' .claude/skills/devlyn:resolve/references/phases/build-gate.md \
+  || ! grep -Fq 'Malformed `state.risk_profile` is also CRITICAL because it can hide enabled risk probes' .agents/skills/devlyn:resolve/references/phases/build-gate.md \
+  || ! grep -Fq 'When `state.risk_profile.risk_probes_enabled == true`, missing `.devlyn/risk-probes.jsonl` is also CRITICAL' config/skills/devlyn:resolve/references/phases/verify.md \
+  || ! grep -Fq 'When `state.risk_profile.risk_probes_enabled == true`, missing `.devlyn/risk-probes.jsonl` is also CRITICAL' .claude/skills/devlyn:resolve/references/phases/verify.md \
+  || ! grep -Fq 'When `state.risk_profile.risk_probes_enabled == true`, missing `.devlyn/risk-probes.jsonl` is also CRITICAL' .agents/skills/devlyn:resolve/references/phases/verify.md; then
+  bad "BUILD_GATE and VERIFY must fail closed when enabled risk probes are missing"
+else
+  ok "BUILD_GATE and VERIFY require enabled risk probes"
+fi
+if grep -Fq 'or any(char.isspace() for char in stripped)' config/skills/_shared/spec-verify-check.py \
+  || grep -Fq 'or any(char.isspace() for char in stripped)' .claude/skills/_shared/spec-verify-check.py \
+  || grep -Fq 'or any(char.isspace() for char in stripped)' .agents/skills/_shared/spec-verify-check.py \
+  || grep -Fq 'or any(char.isspace() for char in stripped)' config/skills/_shared/verify-merge-findings.py \
+  || grep -Fq 'or any(char.isspace() for char in stripped)' .claude/skills/_shared/verify-merge-findings.py \
+  || grep -Fq 'or any(char.isspace() for char in stripped)' .agents/skills/_shared/verify-merge-findings.py \
+  || grep -Fq 'or any(char.isspace() for char in stripped)' benchmark/auto-resolve/scripts/pair_evidence_contract.py; then
+  bad "solo-headroom command detection must not treat descriptive whitespace as a command"
+else
+  ok "solo-headroom command detection rejects descriptive whitespace"
+fi
+if ! grep -Fq '"printf",' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq '"printf",' .claude/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq '"printf",' .agents/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq '"printf",' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq '"printf",' .claude/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq '"printf",' .agents/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq '"printf",' benchmark/auto-resolve/scripts/pair_evidence_contract.py; then
+  bad "solo-headroom command detection must keep explicit printf command support"
+else
+  ok "solo-headroom command detection keeps explicit printf command support"
+fi
+if python3 - <<'PY'
+import ast
+import pathlib
+import sys
+
+files = [
+    pathlib.Path("benchmark/auto-resolve/scripts/pair_evidence_contract.py"),
+    pathlib.Path("config/skills/_shared/spec-verify-check.py"),
+    pathlib.Path(".claude/skills/_shared/spec-verify-check.py"),
+    pathlib.Path(".agents/skills/_shared/spec-verify-check.py"),
+    pathlib.Path("config/skills/_shared/verify-merge-findings.py"),
+    pathlib.Path(".claude/skills/_shared/verify-merge-findings.py"),
+    pathlib.Path(".agents/skills/_shared/verify-merge-findings.py"),
+]
+names = [
+    "COMMAND_PREFIXES",
+    "RESERVED_BACKTICK_TERMS",
+    "OBSERVABLE_COMMAND_MARKERS",
+]
+
+def extract(path: pathlib.Path, name: str) -> tuple[str, ...]:
+    tree = ast.parse(path.read_text(encoding="utf-8"))
+    for node in tree.body:
+        if isinstance(node, ast.Assign):
+            if any(isinstance(target, ast.Name) and target.id == name for target in node.targets):
+                value = ast.literal_eval(node.value)
+                return tuple(sorted(value))
+    raise AssertionError(f"{path}: missing {name}")
+
+baseline = {name: extract(files[0], name) for name in names}
+for path in files[1:]:
+    for name in names:
+        value = extract(path, name)
+        if value != baseline[name]:
+            print(f"{path}: {name} drifted from {files[0]}", file=sys.stderr)
+            print(f"expected={baseline[name]!r}", file=sys.stderr)
+            print(f"actual={value!r}", file=sys.stderr)
+            sys.exit(1)
+PY
+then
+  ok "solo-headroom command detection constants stay in parity"
+else
+  bad "solo-headroom command detection constants must stay in parity"
+fi
+if ! grep -Fq 'required.add("rollback_state")' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'rollback verification text did not require rollback_state probe tag' config/skills/_shared/spec-verify-check.py; then
+  bad "spec-verify-check.py must require rollback_state risk probes for rollback/all-or-nothing verification text"
+fi
+if ! grep -Fq 'spec.expected.json top-level array produced a traceback' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'invalid spec.expected.json produced a traceback' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'NaN spec.expected.json did not report invalid numeric constant' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'NaN risk-probes JSONL did not report invalid numeric constant' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'def reject_json_constant' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'loads_strict_json(line)' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'loads_strict_json(expected_path.read_text())' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'top-level must be a JSON object' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'has invalid JSON' config/skills/_shared/spec-verify-check.py; then
+  bad "spec-verify-check.py self-test must fail malformed spec.expected.json cleanly without traceback"
+fi
+if ! grep -Fq 'required.add("error_contract")' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'error_contract without exit-code evidence was accepted' config/skills/_shared/spec-verify-check.py; then
+  bad "spec-verify-check.py must require error_contract risk probes for invalid/stderr/JSON-error/exit-2 verification text"
+fi
+if ! grep -Fq '"asserts_named_stream_output"' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq '"asserts_error_payload_or_stderr"' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq '"asserts_nonzero_or_exit_2"' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'error_contract without exit-code evidence was accepted' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'stdout_stderr_contract without stream evidence was accepted' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq '`stdout_stderr_contract`: `asserts_named_stream_output`' config/skills/devlyn:resolve/references/phases/probe-derive.md \
+  || ! grep -Fq '`error_contract`: `asserts_error_payload_or_stderr`' config/skills/devlyn:resolve/references/phases/probe-derive.md \
+  || ! grep -Fq '`asserts_nonzero_or_exit_2`' config/skills/devlyn:resolve/SKILL.md; then
+  bad "risk-probe error/stdout-stderr tags must require concrete tag_evidence markers in validator and prompt contract"
+fi
+if ! grep -Fq '"http_error_contract"' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'asserts_http_error_status' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'http error text did not require http_error_contract tag' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'http_error_contract without payload evidence was accepted' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'exact error body shape_contract without exact object evidence was accepted' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'exact error body shape_contract with exact object evidence was rejected' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq '`http_error_contract`: `asserts_http_error_status`' config/skills/devlyn:resolve/references/phases/probe-derive.md \
+  || ! grep -Fq '`http_error_contract` must include `asserts_http_error_status`' config/skills/devlyn:resolve/SKILL.md; then
+  bad "risk-probe HTTP error contracts must require concrete status and payload markers"
+fi
+if ! grep -Fq '"uses_visible_input_key_names"' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq '"asserts_visible_output_key_names"' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq '"asserts_no_unexpected_output_keys"' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'JSON error object text did not require shape_contract tag' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'JSON error object shape_contract with exact object evidence was rejected' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'INLINE_JSON_OBJECT_RE' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'inline JSON object text did not require shape_contract tag' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'inline JSON object shape_contract with key evidence was rejected' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'inline JSON error text did not require shape_contract tag' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'inline JSON error shape_contract with exact object evidence was rejected' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'shape_contract without exact key evidence was accepted' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq '`shape_contract` when the visible text names exact keys' config/skills/devlyn:resolve/references/phases/probe-derive.md \
+  || ! grep -Fq '`shape_contract` must' config/skills/devlyn:resolve/SKILL.md; then
+  bad "risk-probe shape contracts must require exact visible input/output key evidence when visible text names shape"
+fi
+if ! grep -Fq "forbidden[ -]+window" config/skills/_shared/spec-verify-check.py \
+  || grep -Fq "r'blocked|overlap|forbidden|window'" config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'generic forbidden-pattern verification text incorrectly required boundary_overlap' config/skills/_shared/spec-verify-check.py; then
+  bad "risk-probe boundary_overlap must trigger for forbidden windows / blocked overlap, not generic forbidden pattern text"
+fi
+if ! grep -Fq '(?:stock|inventory|balance|availability).{0,80}(?:later|remaining|after failures)' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'stock validation error text incorrectly required prior_consumption' config/skills/_shared/spec-verify-check.py; then
+  bad "risk-probe prior_consumption must trigger on later/remaining state consumption, not plain stock validation errors"
+fi
+if ! grep -Fq '"auth_signature_contract"' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq '"idempotency_replay"' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'asserts_signature_over_exact_bytes' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'duplicate_id_rejected_regardless_of_body' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'webhook signature/replay text did not require auth/idempotency probe tags' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq '`auth_signature_contract`: `asserts_signature_over_exact_bytes`' config/skills/devlyn:resolve/references/phases/probe-derive.md \
+  || ! grep -Fq '`idempotency_replay`: `first_delivery_then_duplicate`' config/skills/devlyn:resolve/references/phases/probe-derive.md \
+  || ! grep -Fq '`auth_signature_contract` must include `asserts_signature_over_exact_bytes`' config/skills/devlyn:resolve/SKILL.md; then
+  bad "risk-probe webhook/signature/replay contracts must require concrete auth_signature_contract and idempotency_replay tags"
+fi
+if ! grep -Fq 'signing|signed' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'same.{0,40}`?id`?' config/skills/_shared/spec-verify-check.py; then
+  bad "risk-probe webhook signature/replay trigger must catch signing/signed and same accepted id wording"
+fi
+if ! grep -Fq 'duplicate[ -]+(?:delivery|event|id)' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'duplicate SKU verification text incorrectly required idempotency_replay' config/skills/_shared/spec-verify-check.py; then
+  bad "risk-probe idempotency_replay must trigger on duplicate delivery/event/id, not duplicate SKU aggregation"
+fi
+if ! grep -Fq '"concurrent_state_consistency"' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'overlapping_mutations_exercised' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'concurrent state text did not require concurrent_state_consistency tag' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq '`concurrent_state_consistency`: `overlapping_mutations_exercised`' config/skills/devlyn:resolve/references/phases/probe-derive.md \
+  || ! grep -Fq '`concurrent_state_consistency` must' config/skills/devlyn:resolve/SKILL.md; then
+  bad "risk-probe concurrent state contracts must require concrete concurrent_state_consistency markers"
+fi
+if ! grep -Fq '"atomic_batch_state"' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'mixed_valid_invalid_batch' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'atomic batch text did not require atomic_batch_state tag' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'atomic_batch_state without success-order evidence was accepted' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq '`atomic_batch_state`: `mixed_valid_invalid_batch`' config/skills/devlyn:resolve/references/phases/probe-derive.md \
+  || ! grep -Fq '`atomic_batch_state` must include `mixed_valid_invalid_batch`' config/skills/devlyn:resolve/SKILL.md; then
+  bad "risk-probe atomic batch contracts must require concrete mixed-failure and success-order markers"
+fi
+
+section "Check 6f: ideate validates sibling spec.expected.json"
+expected_check_missing=0
+for file in \
+  config/skills/devlyn:ideate/SKILL.md \
+  config/skills/devlyn:ideate/references/elicitation.md \
+  config/skills/devlyn:ideate/references/from-spec-mode.md \
+  config/skills/devlyn:ideate/references/project-mode.md \
+  config/skills/devlyn:ideate/references/spec-template.md
+do
+  if ! grep -Fq -- '--check-expected <expected-path>' "$file"; then
+    bad "$file — missing spec.expected.json mechanical validation command"
+    expected_check_missing=1
+  fi
+done
+if [ $expected_check_missing -eq 0 ]; then
+  ok "ideate docs require --check-expected for sibling expected contracts"
+fi
+if ! grep -Fq 'any present actionable solo-headroom hypothesis' config/skills/devlyn:ideate/SKILL.md \
+  || ! grep -Fq 'any present actionable solo-headroom hypothesis' .claude/skills/devlyn:ideate/SKILL.md \
+  || ! grep -Fq 'any present actionable solo-headroom hypothesis' .agents/skills/devlyn:ideate/SKILL.md \
+  || ! grep -Fq 'any present actionable solo-headroom hypothesis' config/skills/devlyn:ideate/references/elicitation.md \
+  || ! grep -Fq 'any present actionable solo-headroom hypothesis' .claude/skills/devlyn:ideate/references/elicitation.md \
+  || ! grep -Fq 'any present actionable solo-headroom hypothesis' .agents/skills/devlyn:ideate/references/elicitation.md \
+  || ! grep -Fq 'any present actionable solo-headroom hypothesis' config/skills/devlyn:ideate/references/project-mode.md \
+  || ! grep -Fq 'any present actionable solo-headroom hypothesis' .claude/skills/devlyn:ideate/references/project-mode.md \
+  || ! grep -Fq 'any present actionable solo-headroom hypothesis' .agents/skills/devlyn:ideate/references/project-mode.md \
+  || ! grep -Fq 'any present actionable solo-headroom hypothesis' config/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'any present actionable solo-headroom hypothesis' .claude/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'any present actionable solo-headroom hypothesis' .agents/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq "legacy inline \`## Verification\` JSON carrier" config/skills/devlyn:ideate/SKILL.md \
+  || ! grep -Fq "legacy inline \`## Verification\` JSON carrier" .claude/skills/devlyn:ideate/SKILL.md \
+  || ! grep -Fq "legacy inline \`## Verification\` JSON carrier" .agents/skills/devlyn:ideate/SKILL.md \
+  || ! grep -Fq "legacy inline \`## Verification\` JSON carrier" config/skills/devlyn:ideate/references/elicitation.md \
+  || ! grep -Fq "legacy inline \`## Verification\` JSON carrier" .claude/skills/devlyn:ideate/references/elicitation.md \
+  || ! grep -Fq "legacy inline \`## Verification\` JSON carrier" .agents/skills/devlyn:ideate/references/elicitation.md \
+  || ! grep -Fq "legacy inline \`## Verification\` JSON carrier" config/skills/devlyn:ideate/references/from-spec-mode.md \
+  || ! grep -Fq "legacy inline \`## Verification\` JSON carrier" .claude/skills/devlyn:ideate/references/from-spec-mode.md \
+  || ! grep -Fq "legacy inline \`## Verification\` JSON carrier" .agents/skills/devlyn:ideate/references/from-spec-mode.md \
+  || ! grep -Fq "inline \`## Verification\` JSON carrier" config/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq "inline \`## Verification\` JSON carrier" .claude/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq "inline \`## Verification\` JSON carrier" .agents/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'spec.expected.json.verification_commands[].cmd' config/skills/devlyn:ideate/SKILL.md \
+  || ! grep -Fq 'spec.expected.json.verification_commands[].cmd' .claude/skills/devlyn:ideate/SKILL.md \
+  || ! grep -Fq 'spec.expected.json.verification_commands[].cmd' .agents/skills/devlyn:ideate/SKILL.md \
+  || ! grep -Fq 'spec.expected.json.verification_commands[].cmd' config/skills/devlyn:ideate/references/elicitation.md \
+  || ! grep -Fq 'spec.expected.json.verification_commands[].cmd' .claude/skills/devlyn:ideate/references/elicitation.md \
+  || ! grep -Fq 'spec.expected.json.verification_commands[].cmd' .agents/skills/devlyn:ideate/references/elicitation.md \
+  || ! grep -Fq 'spec.expected.json.verification_commands[].cmd' config/skills/devlyn:ideate/references/project-mode.md \
+  || ! grep -Fq 'spec.expected.json.verification_commands[].cmd' .claude/skills/devlyn:ideate/references/project-mode.md \
+  || ! grep -Fq 'spec.expected.json.verification_commands[].cmd' .agents/skills/devlyn:ideate/references/project-mode.md \
+  || ! grep -Fq 'spec.expected.json.verification_commands[].cmd' config/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'spec.expected.json.verification_commands[].cmd' .claude/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'spec.expected.json.verification_commands[].cmd' .agents/skills/devlyn:resolve/SKILL.md; then
+  bad "ideate/resolve docs must describe mechanical solo-headroom validation"
+else
+  ok "ideate/resolve docs describe mechanical solo-headroom validation"
+fi
+if ! grep -Fq 'def validate_expected_against_sibling_spec' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'empty verification_commands should fail for runtime specs' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'empty verification_commands should be valid for pure-design specs' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'backticked_observable_miss_commands(spec_text)' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'observable command must match spec.expected.json' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'observable command must match `## Verification` JSON carrier' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'mismatched inline solo-headroom command was accepted by --check' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'matched inline solo-headroom command was rejected by --check' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'mismatched sibling solo-headroom command was accepted by --check-expected' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'matched sibling solo-headroom command was rejected by --check-expected' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'def validate_present_solo_ceiling_avoidance' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'weak solo ceiling avoidance was accepted by --check' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'actionable solo ceiling avoidance was rejected by --check' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'weak sibling solo ceiling avoidance was accepted by --check-expected' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'actionable sibling solo ceiling avoidance was rejected by --check-expected' config/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'backticked_observable_miss_commands(spec_text)' .claude/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'observable command must match spec.expected.json' .claude/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'observable command must match `## Verification` JSON carrier' .claude/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'mismatched inline solo-headroom command was accepted by --check' .claude/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'matched inline solo-headroom command was rejected by --check' .claude/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'mismatched sibling solo-headroom command was accepted by --check-expected' .claude/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'matched sibling solo-headroom command was rejected by --check-expected' .claude/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'def validate_present_solo_ceiling_avoidance' .claude/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'weak solo ceiling avoidance was accepted by --check' .claude/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'actionable solo ceiling avoidance was rejected by --check' .claude/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'weak sibling solo ceiling avoidance was accepted by --check-expected' .claude/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'actionable sibling solo ceiling avoidance was rejected by --check-expected' .claude/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'backticked_observable_miss_commands(spec_text)' .agents/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'observable command must match spec.expected.json' .agents/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'observable command must match `## Verification` JSON carrier' .agents/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'mismatched inline solo-headroom command was accepted by --check' .agents/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'matched inline solo-headroom command was rejected by --check' .agents/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'mismatched sibling solo-headroom command was accepted by --check-expected' .agents/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'matched sibling solo-headroom command was rejected by --check-expected' .agents/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'def validate_present_solo_ceiling_avoidance' .agents/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'weak solo ceiling avoidance was accepted by --check' .agents/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'actionable solo ceiling avoidance was rejected by --check' .agents/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'weak sibling solo ceiling avoidance was accepted by --check-expected' .agents/skills/_shared/spec-verify-check.py \
+  || ! grep -Fq 'actionable sibling solo ceiling avoidance was rejected by --check-expected' .agents/skills/_shared/spec-verify-check.py; then
+  bad "spec-verify-check.py must reject empty expected runtime contracts, weak solo ceiling avoidance, and preserve pure-design escape"
+fi
+if ! grep -Fq 'Verification includes at least one compound scenario that exercises the interaction end-to-end' \
+  config/skills/devlyn:ideate/references/spec-template.md \
+  || ! grep -Fq 'Verification includes at least one compound scenario that exercises the interaction end-to-end' \
+    .claude/skills/devlyn:ideate/references/spec-template.md \
+  || ! grep -Fq 'Verification includes at least one compound scenario that exercises the interaction end-to-end' \
+    .agents/skills/devlyn:ideate/references/spec-template.md; then
+  bad "ideate spec template must require compound interaction verification for pair-relevant high-risk specs"
+else
+  ok "ideate spec template requires compound interaction verification for pair-relevant specs"
+fi
+if ! grep -Fq 'ask for one concrete compound' config/skills/devlyn:ideate/references/elicitation.md \
+  || ! grep -Fq 'ask for one concrete compound' .claude/skills/devlyn:ideate/references/elicitation.md \
+  || ! grep -Fq 'ask for one concrete compound' .agents/skills/devlyn:ideate/references/elicitation.md; then
+  bad "ideate elicitation must ask for compound interaction scenarios when pair-relevant risks appear"
+else
+  ok "ideate elicitation asks for compound interaction scenarios when pair-relevant risks appear"
+fi
+if ! grep -Fq 'solo-headroom hypothesis inside `## Verification`' config/skills/devlyn:ideate/references/spec-template.md \
+  || ! grep -Fq 'solo-headroom hypothesis inside `## Verification`' .claude/skills/devlyn:ideate/references/spec-template.md \
+  || ! grep -Fq 'solo-headroom hypothesis inside `## Verification`' .agents/skills/devlyn:ideate/references/spec-template.md \
+  || ! grep -Fq 'ask for one solo-headroom hypothesis' config/skills/devlyn:ideate/references/elicitation.md \
+  || ! grep -Fq 'ask for one solo-headroom hypothesis' .claude/skills/devlyn:ideate/references/elicitation.md \
+  || ! grep -Fq 'ask for one solo-headroom hypothesis' .agents/skills/devlyn:ideate/references/elicitation.md; then
+  bad "ideate must require a visible solo-headroom hypothesis for benchmark and pair-evidence specs"
+else
+  ok "ideate requires solo-headroom hypothesis for benchmark and pair-evidence specs"
+fi
+if ! grep -Fq 'must literally contain `solo-headroom hypothesis`' config/skills/devlyn:ideate/references/spec-template.md \
+  || ! grep -Fq 'must literally contain `solo-headroom hypothesis`' .claude/skills/devlyn:ideate/references/spec-template.md \
+  || ! grep -Fq 'must literally contain `solo-headroom hypothesis`' .agents/skills/devlyn:ideate/references/spec-template.md \
+  || ! grep -Fq 'backticked line itself must contain `miss`' config/skills/devlyn:ideate/references/spec-template.md \
+  || ! grep -Fq 'backticked line itself must contain `miss`' .claude/skills/devlyn:ideate/references/spec-template.md \
+  || ! grep -Fq 'backticked line itself must contain `miss`' .agents/skills/devlyn:ideate/references/spec-template.md \
+  || ! grep -Fq '`solo_claude`, `miss`, and a backticked observable command' config/skills/devlyn:ideate/references/spec-template.md \
+  || ! grep -Fq '`solo_claude`, `miss`, and a backticked observable command' .claude/skills/devlyn:ideate/references/spec-template.md \
+  || ! grep -Fq '`solo_claude`, `miss`, and a backticked observable command' .agents/skills/devlyn:ideate/references/spec-template.md \
+  || ! grep -Fq 'command/observable' config/skills/devlyn:ideate/references/spec-template.md \
+  || ! grep -Fq 'command/observable' .claude/skills/devlyn:ideate/references/spec-template.md \
+  || ! grep -Fq 'command/observable' .agents/skills/devlyn:ideate/references/spec-template.md \
+  || ! grep -Fq 'must literally contain `solo-headroom hypothesis`' config/skills/devlyn:ideate/references/elicitation.md \
+  || ! grep -Fq 'must literally contain `solo-headroom hypothesis`' .claude/skills/devlyn:ideate/references/elicitation.md \
+  || ! grep -Fq 'must literally contain `solo-headroom hypothesis`' .agents/skills/devlyn:ideate/references/elicitation.md \
+  || ! grep -Fq 'line itself must contain `miss`' config/skills/devlyn:ideate/references/elicitation.md \
+  || ! grep -Fq 'line itself must contain `miss`' .claude/skills/devlyn:ideate/references/elicitation.md \
+  || ! grep -Fq 'line itself must contain `miss`' .agents/skills/devlyn:ideate/references/elicitation.md \
+  || ! grep -Fq 'Do not write a benchmark/risk-probe/pair-evidence spec until this' config/skills/devlyn:ideate/references/elicitation.md \
+  || ! grep -Fq 'Do not write a benchmark/risk-probe/pair-evidence spec until this' .claude/skills/devlyn:ideate/references/elicitation.md \
+  || ! grep -Fq 'Do not write a benchmark/risk-probe/pair-evidence spec until this' .agents/skills/devlyn:ideate/references/elicitation.md \
+  || ! grep -Fq 'spec not ready — solo-headroom hypothesis required' config/skills/devlyn:ideate/references/elicitation.md \
+  || ! grep -Fq 'spec not ready — solo-headroom hypothesis required' .claude/skills/devlyn:ideate/references/elicitation.md \
+  || ! grep -Fq 'spec not ready — solo-headroom hypothesis required' .agents/skills/devlyn:ideate/references/elicitation.md \
+  || ! grep -Fq '`solo_claude`, `miss`, and a backticked observable command' config/skills/devlyn:ideate/references/elicitation.md \
+  || ! grep -Fq '`solo_claude`, `miss`, and a backticked observable command' .claude/skills/devlyn:ideate/references/elicitation.md \
+  || ! grep -Fq '`solo_claude`, `miss`, and a backticked observable command' .agents/skills/devlyn:ideate/references/elicitation.md \
+  || ! grep -Fq 'command/observable' config/skills/devlyn:ideate/references/elicitation.md \
+  || ! grep -Fq 'command/observable' .claude/skills/devlyn:ideate/references/elicitation.md \
+  || ! grep -Fq 'command/observable' .agents/skills/devlyn:ideate/references/elicitation.md \
+  || ! grep -Fq 'Verification literally contains `solo-headroom hypothesis`, `solo_claude`' config/skills/devlyn:ideate/references/from-spec-mode.md \
+  || ! grep -Fq 'Verification literally contains `solo-headroom hypothesis`, `solo_claude`' .claude/skills/devlyn:ideate/references/from-spec-mode.md \
+  || ! grep -Fq 'Verification literally contains `solo-headroom hypothesis`, `solo_claude`' .agents/skills/devlyn:ideate/references/from-spec-mode.md \
+  || ! grep -Fq 'backticked line itself must contain `miss`' config/skills/devlyn:ideate/references/from-spec-mode.md \
+  || ! grep -Fq 'backticked line itself must contain `miss`' .claude/skills/devlyn:ideate/references/from-spec-mode.md \
+  || ! grep -Fq 'backticked line itself must contain `miss`' .agents/skills/devlyn:ideate/references/from-spec-mode.md \
+  || ! grep -Fq 'command/observable' config/skills/devlyn:ideate/references/from-spec-mode.md \
+  || ! grep -Fq 'command/observable' .claude/skills/devlyn:ideate/references/from-spec-mode.md \
+  || ! grep -Fq 'command/observable' .agents/skills/devlyn:ideate/references/from-spec-mode.md \
+  || ! grep -Fq 'spec.expected.json.verification_commands[].cmd' config/skills/devlyn:ideate/references/from-spec-mode.md \
+  || ! grep -Fq 'spec.expected.json.verification_commands[].cmd' .claude/skills/devlyn:ideate/references/from-spec-mode.md \
+  || ! grep -Fq 'spec.expected.json.verification_commands[].cmd' .agents/skills/devlyn:ideate/references/from-spec-mode.md \
+  || ! grep -Fq "source for VERIFY's canonical \`spec.solo_headroom_hypothesis\` trigger reason" config/skills/devlyn:ideate/references/spec-template.md \
+  || ! grep -Fq "source for VERIFY's canonical \`spec.solo_headroom_hypothesis\` trigger reason" .claude/skills/devlyn:ideate/references/spec-template.md \
+  || ! grep -Fq "source for VERIFY's canonical \`spec.solo_headroom_hypothesis\` trigger reason" .agents/skills/devlyn:ideate/references/spec-template.md \
+  || ! grep -Fq 'spec.expected.json.verification_commands[].cmd' config/skills/devlyn:ideate/references/spec-template.md \
+  || ! grep -Fq 'spec.expected.json.verification_commands[].cmd' .claude/skills/devlyn:ideate/references/spec-template.md \
+  || ! grep -Fq 'spec.expected.json.verification_commands[].cmd' .agents/skills/devlyn:ideate/references/spec-template.md; then
+  bad "ideate solo-headroom hypothesis prompt must match the actionable checker contract"
+else
+  ok "ideate solo-headroom hypothesis prompt matches checker contract"
+fi
+if ! grep -Fq 'quick mode must not infer a solo-headroom hypothesis' config/skills/devlyn:ideate/references/elicitation.md \
+  || ! grep -Fq 'quick mode must not infer a solo-headroom hypothesis' .claude/skills/devlyn:ideate/references/elicitation.md \
+  || ! grep -Fq 'quick mode must not infer a solo-headroom hypothesis' .agents/skills/devlyn:ideate/references/elicitation.md \
+  || ! grep -Fq 'do not infer a solo-headroom hypothesis' config/skills/devlyn:ideate/SKILL.md \
+  || ! grep -Fq 'do not infer a solo-headroom hypothesis' .claude/skills/devlyn:ideate/SKILL.md \
+  || ! grep -Fq 'do not infer a solo-headroom hypothesis' .agents/skills/devlyn:ideate/SKILL.md; then
+  bad "ideate quick mode must not invent solo-headroom hypotheses"
+else
+  ok "ideate quick mode does not invent solo-headroom hypotheses"
+fi
+if ! grep -Fq 'solo ceiling avoidance' config/skills/devlyn:ideate/references/spec-template.md \
+  || ! grep -Fq 'solo ceiling avoidance' .claude/skills/devlyn:ideate/references/spec-template.md \
+  || ! grep -Fq 'solo ceiling avoidance' .agents/skills/devlyn:ideate/references/spec-template.md \
+  || ! grep -Fq 'rejected or solo-saturated controls such as `S2`-`S6`' config/skills/devlyn:ideate/references/spec-template.md \
+  || ! grep -Fq 'rejected or solo-saturated controls such as `S2`-`S6`' .claude/skills/devlyn:ideate/references/spec-template.md \
+  || ! grep -Fq 'rejected or solo-saturated controls such as `S2`-`S6`' .agents/skills/devlyn:ideate/references/spec-template.md \
+  || ! grep -Fq 'Solo ceiling avoidance' config/skills/devlyn:ideate/references/elicitation.md \
+  || ! grep -Fq 'Solo ceiling avoidance' .claude/skills/devlyn:ideate/references/elicitation.md \
+  || ! grep -Fq 'Solo ceiling avoidance' .agents/skills/devlyn:ideate/references/elicitation.md \
+  || ! grep -Fq 'spec not ready — solo ceiling avoidance required' config/skills/devlyn:ideate/references/elicitation.md \
+  || ! grep -Fq 'spec not ready — solo ceiling avoidance required' .claude/skills/devlyn:ideate/references/elicitation.md \
+  || ! grep -Fq 'spec not ready — solo ceiling avoidance required' .agents/skills/devlyn:ideate/references/elicitation.md \
+  || ! grep -Fq 'pair-evidence not ready — Pair-candidate headroom is unproven until the spec states solo ceiling avoidance' config/skills/devlyn:ideate/references/from-spec-mode.md \
+  || ! grep -Fq 'pair-evidence not ready — Pair-candidate headroom is unproven until the spec states solo ceiling avoidance' .claude/skills/devlyn:ideate/references/from-spec-mode.md \
+  || ! grep -Fq 'pair-evidence not ready — Pair-candidate headroom is unproven until the spec states solo ceiling avoidance' .agents/skills/devlyn:ideate/references/from-spec-mode.md \
+  || ! grep -Fq 'also do not infer solo ceiling avoidance' config/skills/devlyn:ideate/SKILL.md \
+  || ! grep -Fq 'also do not infer solo ceiling avoidance' .claude/skills/devlyn:ideate/SKILL.md \
+  || ! grep -Fq 'also do not infer solo ceiling avoidance' .agents/skills/devlyn:ideate/SKILL.md \
+  || ! grep -Fq 'per-feature Verification must also include a solo ceiling avoidance note' config/skills/devlyn:ideate/references/project-mode.md \
+  || ! grep -Fq 'per-feature Verification must also include a solo ceiling avoidance note' .claude/skills/devlyn:ideate/references/project-mode.md \
+  || ! grep -Fq 'per-feature Verification must also include a solo ceiling avoidance note' .agents/skills/devlyn:ideate/references/project-mode.md; then
+  bad "ideate must require solo ceiling avoidance for new unmeasured pair candidates"
+else
+  ok "ideate requires solo ceiling avoidance for new unmeasured pair candidates"
+fi
+if ! grep -Fq 'complexity: medium' config/skills/devlyn:ideate/references/spec-template.md \
+  || ! grep -Fq 'complexity: medium' .claude/skills/devlyn:ideate/references/spec-template.md \
+  || ! grep -Fq 'complexity: medium' .agents/skills/devlyn:ideate/references/spec-template.md \
+  || ! grep -Fq 'Complexity signal' config/skills/devlyn:ideate/references/elicitation.md \
+  || ! grep -Fq 'Complexity signal' .claude/skills/devlyn:ideate/references/elicitation.md \
+  || ! grep -Fq 'Complexity signal' .agents/skills/devlyn:ideate/references/elicitation.md \
+  || ! grep -Fq 'downstream VERIFY pair-trigger signal' config/skills/devlyn:ideate/references/elicitation.md \
+  || ! grep -Fq 'downstream VERIFY pair-trigger signal' .claude/skills/devlyn:ideate/references/elicitation.md \
+  || ! grep -Fq 'downstream VERIFY pair-trigger signal' .agents/skills/devlyn:ideate/references/elicitation.md \
+  || ! grep -Fq 'complexity=medium default' config/skills/devlyn:ideate/references/from-spec-mode.md \
+  || ! grep -Fq 'complexity=medium default' .claude/skills/devlyn:ideate/references/from-spec-mode.md \
+  || ! grep -Fq 'complexity=medium default' .agents/skills/devlyn:ideate/references/from-spec-mode.md \
+  || ! grep -Fq 'supported `complexity` frontmatter' config/skills/devlyn:ideate/SKILL.md \
+  || ! grep -Fq 'supported `complexity` frontmatter' .claude/skills/devlyn:ideate/SKILL.md \
+  || ! grep -Fq 'supported `complexity` frontmatter' .agents/skills/devlyn:ideate/SKILL.md \
+  || ! grep -Fq 'supported `complexity` frontmatter' config/skills/devlyn:ideate/references/elicitation.md \
+  || ! grep -Fq 'supported `complexity` frontmatter' .claude/skills/devlyn:ideate/references/elicitation.md \
+  || ! grep -Fq 'supported `complexity` frontmatter' .agents/skills/devlyn:ideate/references/elicitation.md \
+  || ! grep -Fq 'supported `complexity` frontmatter' config/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'supported `complexity` frontmatter' .claude/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'supported `complexity` frontmatter' .agents/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'sibling spec `complexity` frontmatter' config/skills/devlyn:ideate/SKILL.md \
+  || ! grep -Fq 'sibling spec `complexity` frontmatter' .claude/skills/devlyn:ideate/SKILL.md \
+  || ! grep -Fq 'sibling spec `complexity` frontmatter' .agents/skills/devlyn:ideate/SKILL.md \
+  || ! grep -Fq 'sibling spec `complexity` frontmatter' config/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'sibling spec `complexity` frontmatter' .claude/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'sibling spec `complexity` frontmatter' .agents/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'Frontmatter has `id`, `title`, `kind`, `status: planned`, `complexity`' config/skills/devlyn:ideate/SKILL.md \
+  || ! grep -Fq 'Frontmatter has `id`, `title`, `kind`, `status: planned`, `complexity`' .claude/skills/devlyn:ideate/SKILL.md \
+  || ! grep -Fq 'Frontmatter has `id`, `title`, `kind`, `status: planned`, `complexity`' .agents/skills/devlyn:ideate/SKILL.md; then
+  bad "ideate specs must emit complexity frontmatter for resolve pair triggers"
+else
+  ok "ideate specs emit complexity frontmatter for resolve pair triggers"
+fi
+if ! grep -Fq 'warning: Verification may need one compound end-to-end scenario before pair-relevant risks are measurable' \
+  config/skills/devlyn:ideate/references/from-spec-mode.md \
+  || ! grep -Fq 'warning: Verification may need one compound end-to-end scenario before pair-relevant risks are measurable' \
+    .claude/skills/devlyn:ideate/references/from-spec-mode.md \
+  || ! grep -Fq 'warning: Verification may need one compound end-to-end scenario before pair-relevant risks are measurable' \
+    .agents/skills/devlyn:ideate/references/from-spec-mode.md; then
+  bad "ideate from-spec mode must warn when preserved high-risk specs lack compound verification"
+else
+  ok "ideate from-spec mode warns on pair-relevant specs with weak verification"
+fi
+if ! grep -Fq 'pair-evidence not ready — Pair-candidate headroom is unproven until the spec states a solo-headroom hypothesis' \
+  config/skills/devlyn:ideate/references/from-spec-mode.md \
+  || ! grep -Fq 'pair-evidence not ready — Pair-candidate headroom is unproven until the spec states a solo-headroom hypothesis' \
+    .claude/skills/devlyn:ideate/references/from-spec-mode.md \
+  || ! grep -Fq 'pair-evidence not ready — Pair-candidate headroom is unproven until the spec states a solo-headroom hypothesis' \
+    .agents/skills/devlyn:ideate/references/from-spec-mode.md \
+  || ! grep -Fq 'Do not call' config/skills/devlyn:ideate/references/from-spec-mode.md \
+  || ! grep -Fq 'Do not call' .claude/skills/devlyn:ideate/references/from-spec-mode.md \
+  || ! grep -Fq 'Do not call' .agents/skills/devlyn:ideate/references/from-spec-mode.md \
+  || ! grep -Fq 'announcement must say `pair-evidence not ready`' config/skills/devlyn:ideate/SKILL.md \
+  || ! grep -Fq 'announcement must say `pair-evidence not ready`' .claude/skills/devlyn:ideate/SKILL.md \
+  || ! grep -Fq 'announcement must say `pair-evidence not ready`' .agents/skills/devlyn:ideate/SKILL.md; then
+  bad "ideate from-spec mode must warn when pair-candidate specs lack solo-headroom hypothesis"
+else
+  ok "ideate from-spec mode warns on missing solo-headroom hypothesis"
+fi
+if ! grep -Fq 'per-feature Verification must' config/skills/devlyn:ideate/references/project-mode.md \
+  || ! grep -Fq 'per-feature Verification must' .claude/skills/devlyn:ideate/references/project-mode.md \
+  || ! grep -Fq 'per-feature Verification must' .agents/skills/devlyn:ideate/references/project-mode.md; then
+  bad "ideate project mode must require compound verification inside each pair-relevant feature spec"
+else
+  ok "ideate project mode keeps compound verification inside pair-relevant feature specs"
+fi
+if ! grep -Fq 'per-feature Verification must include a solo-headroom' config/skills/devlyn:ideate/references/project-mode.md \
+  || ! grep -Fq 'per-feature Verification must include a solo-headroom' .claude/skills/devlyn:ideate/references/project-mode.md \
+  || ! grep -Fq 'per-feature Verification must include a solo-headroom' .agents/skills/devlyn:ideate/references/project-mode.md \
+  || rg -q 'Context or Verification' config/skills/devlyn:ideate/references .claude/skills/devlyn:ideate/references .agents/skills/devlyn:ideate/references; then
+  bad "ideate project mode must keep solo-headroom hypothesis inside each pair-candidate feature spec"
+else
+  ok "ideate project mode keeps solo-headroom hypothesis inside each pair-candidate feature spec"
+fi
+if ! grep -Fq 'feature spec must literally contain' \
+  config/skills/devlyn:ideate/references/project-mode.md \
+  || ! grep -Fq 'feature spec must literally contain' \
+    .claude/skills/devlyn:ideate/references/project-mode.md \
+  || ! grep -Fq 'feature spec must literally contain' \
+    .agents/skills/devlyn:ideate/references/project-mode.md \
+  || ! grep -Fq '`solo-headroom hypothesis`, `solo_claude`, `miss`, and a backticked' \
+    config/skills/devlyn:ideate/references/project-mode.md \
+  || ! grep -Fq '`solo-headroom hypothesis`, `solo_claude`, `miss`, and a backticked' \
+    .claude/skills/devlyn:ideate/references/project-mode.md \
+  || ! grep -Fq '`solo-headroom hypothesis`, `solo_claude`, `miss`, and a backticked' \
+    .agents/skills/devlyn:ideate/references/project-mode.md \
+  || ! grep -Fq 'behavior a capable' config/skills/devlyn:ideate/references/project-mode.md \
+  || ! grep -Fq 'behavior a capable' .claude/skills/devlyn:ideate/references/project-mode.md \
+  || ! grep -Fq 'behavior a capable' .agents/skills/devlyn:ideate/references/project-mode.md \
+  || ! grep -Fq 'backticked line itself must' config/skills/devlyn:ideate/references/project-mode.md \
+  || ! grep -Fq 'backticked line itself must' .claude/skills/devlyn:ideate/references/project-mode.md \
+  || ! grep -Fq 'backticked line itself must' .agents/skills/devlyn:ideate/references/project-mode.md \
+  || ! grep -Fq 'command/observable' \
+    config/skills/devlyn:ideate/references/project-mode.md \
+  || ! grep -Fq 'command/observable' \
+    .claude/skills/devlyn:ideate/references/project-mode.md \
+  || ! grep -Fq 'command/observable' \
+    .agents/skills/devlyn:ideate/references/project-mode.md; then
+  bad "ideate project mode solo-headroom prompt must keep the actionable checker contract"
+else
+  ok "ideate project mode solo-headroom prompt keeps checker contract"
+fi
+
+if ! grep -Fq 'If the visible spec includes a solo-headroom hypothesis, the first probe must' \
+  config/skills/devlyn:resolve/references/phases/probe-derive.md \
+  || ! grep -Fq 'If the visible spec includes a solo-headroom hypothesis, the first probe must' \
+    .claude/skills/devlyn:resolve/references/phases/probe-derive.md \
+  || ! grep -Fq 'If the visible spec includes a solo-headroom hypothesis, the first probe must' \
+    .agents/skills/devlyn:resolve/references/phases/probe-derive.md \
+  || ! grep -Fq 'When the visible spec includes a solo-headroom hypothesis, the first probe must' \
+    config/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'When the visible spec includes a solo-headroom hypothesis, the first probe must' \
+    .claude/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'When the visible spec includes a solo-headroom hypothesis, the first probe must' \
+    .agents/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'its `cmd` must contain the hypothesis'\''s backticked' \
+    config/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'its `cmd` must contain the hypothesis'\''s backticked' \
+    .claude/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'its `cmd` must contain the hypothesis'\''s backticked' \
+    .agents/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'its `derived_from` must reference the hypothesis bullet' \
+    config/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'its `derived_from` must reference the hypothesis bullet' \
+    .claude/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'its `derived_from` must reference the hypothesis bullet' \
+    .agents/skills/devlyn:resolve/SKILL.md; then
+  bad "resolve risk-probe prompts must consume solo-headroom hypothesis before pair-evidence work"
+else
+  ok "resolve risk-probe prompts consume solo-headroom hypothesis"
+fi
+if ! grep -Fq 'the behavior the spec says `solo_claude` is expected to miss' \
+  config/skills/devlyn:resolve/references/phases/probe-derive.md \
+  || ! grep -Fq 'the behavior the spec says `solo_claude` is expected to miss' \
+    .claude/skills/devlyn:resolve/references/phases/probe-derive.md \
+  || ! grep -Fq 'the behavior the spec says `solo_claude` is expected to miss' \
+    .agents/skills/devlyn:resolve/references/phases/probe-derive.md \
+  || ! grep -Fq 'exercise the stated `solo_claude` miss' \
+    config/skills/devlyn:resolve/references/phases/probe-derive.md \
+  || ! grep -Fq 'exercise the stated `solo_claude` miss' \
+    .claude/skills/devlyn:resolve/references/phases/probe-derive.md \
+  || ! grep -Fq 'exercise the stated `solo_claude` miss' \
+    .agents/skills/devlyn:resolve/references/phases/probe-derive.md \
+  || ! grep -Fq 'with a `cmd` containing the hypothesis'\''s' \
+    config/skills/devlyn:resolve/references/phases/probe-derive.md \
+  || ! grep -Fq 'with a `cmd` containing the hypothesis'\''s' \
+    .claude/skills/devlyn:resolve/references/phases/probe-derive.md \
+  || ! grep -Fq 'with a `cmd` containing the hypothesis'\''s' \
+    .agents/skills/devlyn:resolve/references/phases/probe-derive.md \
+  || ! grep -Fq '`derived_from` pointing at the hypothesis' \
+    config/skills/devlyn:resolve/references/phases/probe-derive.md \
+  || ! grep -Fq '`derived_from` pointing at the hypothesis' \
+    .claude/skills/devlyn:resolve/references/phases/probe-derive.md \
+  || ! grep -Fq '`derived_from` pointing at the hypothesis' \
+    .agents/skills/devlyn:resolve/references/phases/probe-derive.md; then
+  bad "resolve risk-probe solo-headroom prompt must target the stated solo_claude miss"
+else
+  ok "resolve risk-probe solo-headroom prompt targets the stated solo_claude miss"
+fi
+if ! grep -Fq 'If the spec includes a solo-headroom hypothesis, one of the two targeted' \
+  config/skills/devlyn:resolve/references/phases/verify.md \
+  || ! grep -Fq 'If the spec includes a solo-headroom hypothesis, one of the two targeted' \
+    .claude/skills/devlyn:resolve/references/phases/verify.md \
+  || ! grep -Fq 'If the spec includes a solo-headroom hypothesis, one of the two targeted' \
+    .agents/skills/devlyn:resolve/references/phases/verify.md \
+  || ! grep -Fq 'If the spec includes a solo-headroom hypothesis, one of those targeted probes must' \
+    config/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'If the spec includes a solo-headroom hypothesis, one of those targeted probes must' \
+    .claude/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'If the spec includes a solo-headroom hypothesis, one of those targeted probes must' \
+    .agents/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'using the hypothesis'\''s backticked observable command as its command anchor' \
+    config/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'using the hypothesis'\''s backticked observable command as its command anchor' \
+    .claude/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'using the hypothesis'\''s backticked observable command as its command anchor' \
+    .agents/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'hypothesis'\''s backticked observable command as its command anchor' \
+    config/skills/devlyn:resolve/references/phases/verify.md \
+  || ! grep -Fq 'hypothesis'\''s backticked observable command as its command anchor' \
+    .claude/skills/devlyn:resolve/references/phases/verify.md \
+  || ! grep -Fq 'hypothesis'\''s backticked observable command as its command anchor' \
+    .agents/skills/devlyn:resolve/references/phases/verify.md; then
+  bad "resolve pair-JUDGE prompts must prioritize solo-headroom hypothesis"
 else
-  bad "spec-verify-check.py risk-probe self-test failed"
+  ok "resolve pair-JUDGE prompts prioritize solo-headroom hypothesis"
+fi
+if ! grep -Fq 'current free-form `state.complexity` is `"large"`' config/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'current free-form `state.complexity` is `"large"`' .claude/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'current free-form `state.complexity` is `"large"`' .agents/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'legacy/external spec `complexity: large` is accepted for compatibility' config/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'legacy/external spec `complexity: large` is accepted for compatibility' .claude/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'legacy/external spec `complexity: large` is accepted for compatibility' .agents/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'current free-form `state.complexity` of `"large"`' config/skills/devlyn:resolve/references/state-schema.md \
+  || ! grep -Fq 'current free-form `state.complexity` of `"large"`' .claude/skills/devlyn:resolve/references/state-schema.md \
+  || ! grep -Fq 'current free-form `state.complexity` of `"large"`' .agents/skills/devlyn:resolve/references/state-schema.md \
+  || ! grep -Fq 'Legacy/external spec `complexity: large` remains accepted for compatibility' config/skills/devlyn:resolve/references/state-schema.md \
+  || ! grep -Fq 'Legacy/external spec `complexity: large` remains accepted for compatibility' .claude/skills/devlyn:resolve/references/state-schema.md \
+  || ! grep -Fq 'Legacy/external spec `complexity: large` remains accepted for compatibility' .agents/skills/devlyn:resolve/references/state-schema.md \
+  || ! grep -Fq 'Current free-form `state.complexity` is `"large"`' config/skills/devlyn:resolve/references/phases/verify.md \
+  || ! grep -Fq 'Current free-form `state.complexity` is `"large"`' .claude/skills/devlyn:resolve/references/phases/verify.md \
+  || ! grep -Fq 'Current free-form `state.complexity` is `"large"`' .agents/skills/devlyn:resolve/references/phases/verify.md \
+  || ! grep -Fq 'legacy/external spec' config/skills/devlyn:resolve/references/phases/verify.md \
+  || ! grep -Fq 'legacy/external spec' .claude/skills/devlyn:resolve/references/phases/verify.md \
+  || ! grep -Fq 'legacy/external spec' .agents/skills/devlyn:resolve/references/phases/verify.md; then
+  bad "resolve VERIFY docs must distinguish current large complexity, legacy high state, and legacy large spec compatibility"
+else
+  ok "resolve VERIFY docs distinguish current large complexity, legacy high state, and legacy large spec compatibility"
+fi
+if ! grep -Fq 'def spec_has_solo_headroom_hypothesis' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'spec.solo_headroom_hypothesis' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'and "solo_claude" in lower' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'and "miss" in lower' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'OBSERVABLE_COMMAND_MARKERS = ("command", "observable", "expose")' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'def is_command_like_backtick' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'for key in ("spec_path", "criteria_path")' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'has_backticked_observable_command(text)' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'Observable command: `node check.js` exposes behavior' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'observable `SOLO_CLAUDE` exposes the miss' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'observable `priority rollback` exposes the miss' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'implementation token `rollback`' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq '`SOLO_CLAUDE` should miss' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq '{"source": {"criteria_path": str(criteria_path)}}' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'spec_has_solo_headroom_hypothesis(' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq ') is False' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq ') is True' config/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'def spec_has_solo_headroom_hypothesis' .claude/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'spec.solo_headroom_hypothesis' .claude/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'and "solo_claude" in lower' .claude/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'and "miss" in lower' .claude/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'OBSERVABLE_COMMAND_MARKERS = ("command", "observable", "expose")' .claude/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'def is_command_like_backtick' .claude/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'for key in ("spec_path", "criteria_path")' .claude/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'has_backticked_observable_command(text)' .claude/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'Observable command: `node check.js` exposes behavior' .claude/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'observable `SOLO_CLAUDE` exposes the miss' .claude/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'observable `priority rollback` exposes the miss' .claude/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'implementation token `rollback`' .claude/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq '`SOLO_CLAUDE` should miss' .claude/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq '{"source": {"criteria_path": str(criteria_path)}}' .claude/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'spec_has_solo_headroom_hypothesis(' .claude/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq ') is False' .claude/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq ') is True' .claude/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'def spec_has_solo_headroom_hypothesis' .agents/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'spec.solo_headroom_hypothesis' .agents/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'and "solo_claude" in lower' .agents/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'and "miss" in lower' .agents/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'OBSERVABLE_COMMAND_MARKERS = ("command", "observable", "expose")' .agents/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'def is_command_like_backtick' .agents/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'for key in ("spec_path", "criteria_path")' .agents/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'has_backticked_observable_command(text)' .agents/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'Observable command: `node check.js` exposes behavior' .agents/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'observable `SOLO_CLAUDE` exposes the miss' .agents/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'observable `priority rollback` exposes the miss' .agents/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'implementation token `rollback`' .agents/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq '`SOLO_CLAUDE` should miss' .agents/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq '{"source": {"criteria_path": str(criteria_path)}}' .agents/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'spec_has_solo_headroom_hypothesis(' .agents/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq ') is False' .agents/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq ') is True' .agents/skills/_shared/verify-merge-findings.py \
+  || ! grep -Fq 'The spec includes an actionable solo-headroom hypothesis.' \
+    config/skills/devlyn:resolve/references/phases/verify.md \
+  || ! grep -Fq 'The spec includes an actionable solo-headroom hypothesis.' \
+    .claude/skills/devlyn:resolve/references/phases/verify.md \
+  || ! grep -Fq 'The spec includes an actionable solo-headroom hypothesis.' \
+    .agents/skills/devlyn:resolve/references/phases/verify.md \
+  || ! grep -Fq 'actionable solo-headroom hypotheses' config/skills/devlyn:resolve/references/state-schema.md \
+  || ! grep -Fq 'actionable solo-headroom hypotheses' .claude/skills/devlyn:resolve/references/state-schema.md \
+  || ! grep -Fq 'actionable solo-headroom hypotheses' .agents/skills/devlyn:resolve/references/state-schema.md \
+  || ! grep -Fq 'same actionable solo-headroom hypothesis is a VERIFY pair-trigger reason' config/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'same actionable solo-headroom hypothesis is a VERIFY pair-trigger reason' .claude/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'same actionable solo-headroom hypothesis is a VERIFY pair-trigger reason' .agents/skills/devlyn:resolve/SKILL.md; then
+  bad "resolve VERIFY pair trigger must include actionable solo-headroom hypothesis specs"
+else
+  ok "resolve VERIFY pair trigger includes actionable solo-headroom hypothesis specs"
+fi
+if ! grep -Fq 'pair_evidence_intent' config/skills/devlyn:resolve/references/free-form-mode.md \
+  || ! grep -Fq 'pair_evidence_intent' .claude/skills/devlyn:resolve/references/free-form-mode.md \
+  || ! grep -Fq 'pair_evidence_intent' .agents/skills/devlyn:resolve/references/free-form-mode.md \
+  || ! grep -Fq 'has_actionable_solo_headroom' config/skills/devlyn:resolve/references/free-form-mode.md \
+  || ! grep -Fq 'has_actionable_solo_headroom' .claude/skills/devlyn:resolve/references/free-form-mode.md \
+  || ! grep -Fq 'has_actionable_solo_headroom' .agents/skills/devlyn:resolve/references/free-form-mode.md \
+  || ! grep -Fq 'state.source.type = "generated"' config/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'state.source.type = "generated"' .claude/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'state.source.type = "generated"' .agents/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'state.source.criteria_path = ".devlyn/criteria.generated.md"' config/skills/devlyn:resolve/references/free-form-mode.md \
+  || ! grep -Fq 'state.source.criteria_path = ".devlyn/criteria.generated.md"' .claude/skills/devlyn:resolve/references/free-form-mode.md \
+  || ! grep -Fq 'state.source.criteria_path = ".devlyn/criteria.generated.md"' .agents/skills/devlyn:resolve/references/free-form-mode.md \
+  || ! grep -Fq 'state.source.criteria_sha256' config/skills/devlyn:resolve/references/free-form-mode.md \
+  || ! grep -Fq 'state.source.criteria_sha256' .claude/skills/devlyn:resolve/references/free-form-mode.md \
+  || ! grep -Fq 'state.source.criteria_sha256' .agents/skills/devlyn:resolve/references/free-form-mode.md \
+  || ! grep -Fq 'state.source.criteria_sha256` for generated free-form mode' config/skills/devlyn:resolve/references/phases/verify.md \
+  || ! grep -Fq 'state.source.criteria_sha256` for generated free-form mode' .claude/skills/devlyn:resolve/references/phases/verify.md \
+  || ! grep -Fq 'state.source.criteria_sha256` for generated free-form mode' .agents/skills/devlyn:resolve/references/phases/verify.md \
+  || ! grep -Fq 'Free-form mode sets `type: "generated"`' config/skills/devlyn:resolve/references/state-schema.md \
+  || ! grep -Fq 'Free-form mode sets `type: "generated"`' .claude/skills/devlyn:resolve/references/state-schema.md \
+  || ! grep -Fq 'Free-form mode sets `type: "generated"`' .agents/skills/devlyn:resolve/references/state-schema.md \
+  || ! grep -Fq 'backticked observable command line that itself contains `miss`' config/skills/devlyn:resolve/references/free-form-mode.md \
+  || ! grep -Fq 'backticked observable command line that itself contains `miss`' .claude/skills/devlyn:resolve/references/free-form-mode.md \
+  || ! grep -Fq 'backticked observable command line that itself contains `miss`' .agents/skills/devlyn:resolve/references/free-form-mode.md \
+  || ! grep -Fq 'BLOCKED:solo-headroom-hypothesis-required' config/skills/devlyn:resolve/references/free-form-mode.md \
+  || ! grep -Fq 'BLOCKED:solo-headroom-hypothesis-required' .claude/skills/devlyn:resolve/references/free-form-mode.md \
+  || ! grep -Fq 'BLOCKED:solo-headroom-hypothesis-required' .agents/skills/devlyn:resolve/references/free-form-mode.md \
+  || ! grep -Fq 'pair-evidence intent without an actionable solo-headroom hypothesis must halt' config/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'pair-evidence intent without an actionable solo-headroom hypothesis must halt' .claude/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'pair-evidence intent without an actionable solo-headroom hypothesis must halt' .agents/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'preserve that literal hypothesis in `.devlyn/criteria.generated.md`' config/skills/devlyn:resolve/references/free-form-mode.md \
+  || ! grep -Fq 'preserve that literal hypothesis in `.devlyn/criteria.generated.md`' .claude/skills/devlyn:resolve/references/free-form-mode.md \
+  || ! grep -Fq 'preserve that literal hypothesis in `.devlyn/criteria.generated.md`' .agents/skills/devlyn:resolve/references/free-form-mode.md \
+  || ! grep -Fq 'emit the canonical `spec.solo_headroom_hypothesis` pair trigger reason' config/skills/devlyn:resolve/references/free-form-mode.md \
+  || ! grep -Fq 'emit the canonical `spec.solo_headroom_hypothesis` pair trigger reason' .claude/skills/devlyn:resolve/references/free-form-mode.md \
+  || ! grep -Fq 'emit the canonical `spec.solo_headroom_hypothesis` pair trigger reason' .agents/skills/devlyn:resolve/references/free-form-mode.md \
+  || ! grep -Fq '/devlyn:ideate` guidance for `BLOCKED:solo-headroom-hypothesis-required`' config/skills/devlyn:resolve/references/state-schema.md \
+  || ! grep -Fq '/devlyn:ideate` guidance for `BLOCKED:solo-headroom-hypothesis-required`' .claude/skills/devlyn:resolve/references/state-schema.md \
+  || ! grep -Fq '/devlyn:ideate` guidance for `BLOCKED:solo-headroom-hypothesis-required`' .agents/skills/devlyn:resolve/references/state-schema.md \
+  || ! grep -Fq '/devlyn:ideate` guidance after `BLOCKED:solo-headroom-hypothesis-required`' config/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq '/devlyn:ideate` guidance after `BLOCKED:solo-headroom-hypothesis-required`' .claude/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq '/devlyn:ideate` guidance after `BLOCKED:solo-headroom-hypothesis-required`' .agents/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'Free-form goals that ask for benchmark evidence, pair-evidence, risk-probe' README.md \
+  || ! grep -Fq '`/devlyn:resolve` stops with `BLOCKED:solo-headroom-hypothesis-required`' README.md \
+  || ! grep -Fq 'points you to `/devlyn:ideate` instead of inventing a weak hypothesis' README.md; then
+  bad "resolve free-form mode must block pair-evidence goals without actionable solo-headroom hypothesis"
+else
+  ok "resolve free-form mode blocks pair-evidence goals without actionable solo-headroom hypothesis"
+fi
+if ! grep -Fq 'unmeasured_pair_candidate_intent' config/skills/devlyn:resolve/references/free-form-mode.md \
+  || ! grep -Fq 'unmeasured_pair_candidate_intent' .claude/skills/devlyn:resolve/references/free-form-mode.md \
+  || ! grep -Fq 'unmeasured_pair_candidate_intent' .agents/skills/devlyn:resolve/references/free-form-mode.md \
+  || ! grep -Fq 'has_solo_ceiling_avoidance' config/skills/devlyn:resolve/references/free-form-mode.md \
+  || ! grep -Fq 'has_solo_ceiling_avoidance' .claude/skills/devlyn:resolve/references/free-form-mode.md \
+  || ! grep -Fq 'has_solo_ceiling_avoidance' .agents/skills/devlyn:resolve/references/free-form-mode.md \
+  || ! grep -Fq 'BLOCKED:solo-ceiling-avoidance-required' config/skills/devlyn:resolve/references/free-form-mode.md \
+  || ! grep -Fq 'BLOCKED:solo-ceiling-avoidance-required' .claude/skills/devlyn:resolve/references/free-form-mode.md \
+  || ! grep -Fq 'BLOCKED:solo-ceiling-avoidance-required' .agents/skills/devlyn:resolve/references/free-form-mode.md \
+  || ! grep -Fq 'unmeasured pair-candidate intent without solo ceiling avoidance must halt' config/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'unmeasured pair-candidate intent without solo ceiling avoidance must halt' .claude/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq 'unmeasured pair-candidate intent without solo ceiling avoidance must halt' .agents/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq '/devlyn:ideate` guidance after `BLOCKED:solo-ceiling-avoidance-required`' config/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq '/devlyn:ideate` guidance after `BLOCKED:solo-ceiling-avoidance-required`' .claude/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq '/devlyn:ideate` guidance after `BLOCKED:solo-ceiling-avoidance-required`' .agents/skills/devlyn:resolve/SKILL.md \
+  || ! grep -Fq '/devlyn:ideate` guidance for `BLOCKED:solo-ceiling-avoidance-required`' config/skills/devlyn:resolve/references/state-schema.md \
+  || ! grep -Fq '/devlyn:ideate` guidance for `BLOCKED:solo-ceiling-avoidance-required`' .claude/skills/devlyn:resolve/references/state-schema.md \
+  || ! grep -Fq '/devlyn:ideate` guidance for `BLOCKED:solo-ceiling-avoidance-required`' .agents/skills/devlyn:resolve/references/state-schema.md \
+  || ! grep -Fq '`/devlyn:resolve` stops with `BLOCKED:solo-ceiling-avoidance-required`' README.md; then
+  bad "resolve free-form mode must block new unmeasured pair candidates without solo ceiling avoidance"
+else
+  ok "resolve free-form mode blocks new unmeasured pair candidates without solo ceiling avoidance"
+fi
+
+section "Check 6g: resolve consumes sibling spec.expected.json"
+sibling_consume_missing=0
+for file in \
+  config/skills/devlyn:resolve/SKILL.md \
+  config/skills/devlyn:resolve/references/phases/build-gate.md \
+  config/skills/devlyn:resolve/references/phases/verify.md
+do
+  if ! grep -Fq 'sibling `spec.expected.json`' "$file"; then
+    bad "$file — missing sibling spec.expected.json consumption contract"
+    sibling_consume_missing=1
+  fi
+done
+for pattern in \
+  'def stage_from_expected' \
+  'stage_from_expected(' \
+  'expected_found, expected_staged, expected_error, expected_path' \
+  'def expected_contract_findings' \
+  'correctness.forbidden-pattern' \
+  'scope.max-deps-added-exceeded' \
+  'SPEC_VERIFY_FINDINGS_FILE'
+do
+  if ! grep -Fq "$pattern" config/skills/_shared/spec-verify-check.py; then
+    bad "spec-verify-check.py missing sibling expected staging implementation: $pattern"
+    sibling_consume_missing=1
+  fi
+done
+if [ $sibling_consume_missing -eq 0 ]; then
+  ok "resolve self-stages and mechanically checks sibling spec.expected.json"
+fi
+
+section "Check 6i: VERIFY mechanical findings are merge-visible"
+verify_mech_missing=0
+for pattern in \
+  'SPEC_VERIFY_PHASE=verify_mechanical' \
+  'SPEC_VERIFY_FINDINGS_FILE=verify-mechanical.findings.jsonl' \
+  'SPEC_VERIFY_FINDING_PREFIX=VERIFY-MECH'
+do
+  if ! grep -Fq "$pattern" config/skills/devlyn:resolve/SKILL.md \
+     || ! grep -Fq "$pattern" config/skills/devlyn:resolve/references/phases/verify.md \
+     || ! grep -Fq "$pattern" config/skills/_shared/spec-verify-check.py; then
+    bad "VERIFY mechanical output contract missing: $pattern"
+    verify_mech_missing=1
+  fi
+done
+if ! grep -Fq '("mechanical", "verify-mechanical.findings.jsonl")' \
+  config/skills/_shared/verify-merge-findings.py; then
+  bad "verify-merge-findings.py does not consume verify-mechanical.findings.jsonl"
+  verify_mech_missing=1
+fi
+if grep -Fq 'All paths emit a CRITICAL finding to' config/skills/_shared/spec-verify-check.py; then
+  bad "spec-verify-check.py has stale single-output CRITICAL finding wording"
+  verify_mech_missing=1
+fi
+if [ $verify_mech_missing -eq 0 ]; then
+  ok "spec-verify VERIFY output routes into verify-merge-findings.py"
+fi
+
+section "Check 6j: VERIFY pair trigger runs after primary JUDGE"
+pair_trigger_order_missing=0
+for file in \
+  config/skills/devlyn:resolve/SKILL.md \
+  .claude/skills/devlyn:resolve/SKILL.md \
+  .agents/skills/devlyn:resolve/SKILL.md
+do
+  if ! grep -Fq 'Pair-mode (cross-model JUDGE) is eligible only after MECHANICAL and the primary JUDGE have no verdict-binding findings' "$file" \
+    || ! grep -Fq 'After MECHANICAL and the primary JUDGE finish, compute `pair_trigger' "$file" \
+    || ! grep -Fq '`risk_profile` is strict typed state' "$file" \
+    || ! grep -Fq 'malformed `risk_profile` is also a VERIFY contract violation' "$file" \
+    || ! grep -Fq 'If MECHANICAL or the primary JUDGE has a verdict-binding finding' "$file"; then
+    bad "$file — VERIFY pair trigger order must be after primary JUDGE"
+    pair_trigger_order_missing=1
+  fi
+done
+for file in \
+  config/skills/devlyn:resolve/references/phases/verify.md \
+  .claude/skills/devlyn:resolve/references/phases/verify.md \
+  .agents/skills/devlyn:resolve/references/phases/verify.md
+do
+  if ! grep -Fq 'Pair-mode is eligible only after MECHANICAL and the primary JUDGE have no' "$file" \
+    || ! grep -Fq 'After MECHANICAL and the primary JUDGE finish, compute and persist this before' "$file" \
+    || ! grep -Fq 'Malformed `state.risk_profile` is a VERIFY contract violation' "$file" \
+    || ! grep -Fq 'primary_judge_blocker' "$file"; then
+    bad "$file — VERIFY phase body must compute pair_trigger after primary JUDGE"
+    pair_trigger_order_missing=1
+  fi
+done
+for file in \
+  config/skills/devlyn:resolve/references/state-schema.md \
+  .claude/skills/devlyn:resolve/references/state-schema.md \
+  .agents/skills/devlyn:resolve/references/state-schema.md
+do
+  if ! grep -Fq 'MECHANICAL and the primary JUDGE have no verdict-binding blockers' "$file" \
+    || ! grep -Fq 'may set only `user_no_pair`, `mechanical_blocker`, `primary_judge_blocker`, or null' "$file" \
+    || ! grep -Fq '`risk_profile` must remain an object with boolean' "$file" \
+    || ! grep -Fq 'state implies a pair decision is required but `pair_trigger` is missing' "$file"; then
+    bad "$file — state schema must document pair_trigger blocker and missing-trigger enforcement"
+    pair_trigger_order_missing=1
+  fi
+done
+if [ $pair_trigger_order_missing -eq 0 ]; then
+  ok "VERIFY pair trigger order waits for primary JUDGE evidence"
+fi
+
+section "Check 6h: No undocumented spec.expected.json.browser_flows field"
+browser_flow_refs=$(grep -RInF 'spec.expected.json.browser_flows' \
+  config/skills README.md bin/ package.json 2>/dev/null || true)
+if [ -z "$browser_flow_refs" ]; then
+  ok "active docs do not advertise unsupported browser_flows schema field"
+else
+  while IFS= read -r f; do bad "$f"; done <<< "$browser_flow_refs"
 fi
 
 section "Check 6e: All-or-nothing probes prove mutable rollback"
@@ -283,22 +1456,2313 @@ else
 fi
 
 # ---------------------------------------------------------------------------
-# 9. Engine-downgrade string is canonical (codex-unavailable, not codex-ping failed).
+# 10a. Bounded Codex probe/judge calls must run isolated.
+#      Pair/risk-probe paths are measured read-only judges, not implementation
+#      phases. They must not inherit user config, AGENTS.md, hooks, pyx-memory,
+#      or other local rules that can add hidden context or transcript side
+#      effects. The wrapper owns the flag expansion; skill docs own requiring
+#      CODEX_MONITORED_ISOLATED=1 for probe-derive and pair-JUDGE.
+# ---------------------------------------------------------------------------
+section "Check 10a: Bounded Codex calls use isolated wrapper mode"
+isolation_missing=0
+for needle in \
+  'CODEX_MONITORED_ISOLATED=1 bash "$CODEX_MONITORED_PATH"' \
+  'CODEX_MONITORED_ISOLATED=1` and `-c model_reasoning_effort=medium' \
+  'CODEX_MONITORED_ISOLATED=1 bash .claude/skills/_shared/codex-monitored.sh'
+do
+  if ! grep -RInF "$needle" config/skills >/dev/null 2>&1; then
+    bad "missing isolated Codex invocation contract: $needle"
+    isolation_missing=1
+  fi
+done
+for flag in \
+  '--ignore-user-config' \
+  '--ignore-rules' \
+  '--ephemeral' \
+  '--disable codex_hooks' \
+  '--disable hooks'
+do
+  if ! grep -F -- "$flag" config/skills/_shared/codex-monitored.sh >/dev/null 2>&1; then
+    bad "codex-monitored.sh missing isolation flag expansion: $flag"
+    isolation_missing=1
+  fi
+done
+for helper in require_positive_int require_nonnegative_int; do
+  if ! grep -F "$helper" config/skills/_shared/codex-monitored.sh >/dev/null 2>&1; then
+    bad "codex-monitored.sh missing numeric env validator: $helper"
+    isolation_missing=1
+  fi
+done
+
+if make_temp_dir tmp_env /tmp/codex-monitored-env.XXXXXX; then
+  if CODEX_MONITORED_HEARTBEAT=0 CODEX_BIN=/bin/true \
+    bash config/skills/_shared/codex-monitored.sh prompt \
+    >"$tmp_env/heartbeat.stdout" 2>"$tmp_env/heartbeat.stderr"; then
+    bad "codex-monitored.sh accepted CODEX_MONITORED_HEARTBEAT=0"
+    isolation_missing=1
+  elif ! grep -F 'CODEX_MONITORED_HEARTBEAT must be > 0' "$tmp_env/heartbeat.stderr" >/dev/null 2>&1; then
+    bad "codex-monitored.sh heartbeat validation emitted wrong error"
+    isolation_missing=1
+  fi
+  if CODEX_MONITORED_TIMEOUT_SEC=abc CODEX_BIN=/bin/true \
+    bash config/skills/_shared/codex-monitored.sh prompt \
+    >"$tmp_env/timeout.stdout" 2>"$tmp_env/timeout.stderr"; then
+    bad "codex-monitored.sh accepted non-numeric CODEX_MONITORED_TIMEOUT_SEC"
+    isolation_missing=1
+  elif ! grep -F 'CODEX_MONITORED_TIMEOUT_SEC must be a non-negative integer' "$tmp_env/timeout.stderr" >/dev/null 2>&1; then
+    bad "codex-monitored.sh timeout validation emitted wrong error"
+    isolation_missing=1
+  fi
+  rm -rf "$tmp_env"
+else
+  isolation_missing=1
+fi
+
+if make_temp_dir tmp_iso /tmp/codex-monitored-isolated.XXXXXX; then
+cat > "$tmp_iso/codex" <<'EOF'
+#!/usr/bin/env bash
+printf '%s\n' "$@" > "$CODEX_FAKE_ARGS_OUT"
+EOF
+  chmod +x "$tmp_iso/codex"
+  CODEX_FAKE_ARGS_OUT="$tmp_iso/args.txt" \
+  CODEX_MONITORED_ISOLATED=1 \
+  CODEX_MONITORED_HEARTBEAT=999 \
+  CODEX_BIN="$tmp_iso/codex" \
+    bash config/skills/_shared/codex-monitored.sh -s read-only prompt \
+    >"$tmp_iso/stdout.txt" 2>"$tmp_iso/stderr.txt"
+  iso_exit=$?
+  if [ $iso_exit -ne 0 ]; then
+    bad "codex-monitored.sh isolated fake invocation exited $iso_exit"
+    isolation_missing=1
+  else
+    for expected in exec --ignore-user-config --ignore-rules --ephemeral --disable codex_hooks --disable hooks -s read-only prompt; do
+      if ! grep -Fx -- "$expected" "$tmp_iso/args.txt" >/dev/null 2>&1; then
+        bad "codex-monitored.sh isolated fake invocation missing arg: $expected"
+        isolation_missing=1
+      fi
+    done
+    if ! grep -F '[codex-monitored] isolated=1' "$tmp_iso/stderr.txt" >/dev/null 2>&1; then
+      bad "codex-monitored.sh isolated fake invocation missing lifecycle marker"
+      isolation_missing=1
+    fi
+  fi
+  rm -rf "$tmp_iso"
+else
+  isolation_missing=1
+fi
+if [ $isolation_missing -eq 0 ]; then
+  ok "bounded Codex probe/judge calls require isolated wrapper mode"
+fi
+
+# ---------------------------------------------------------------------------
+# 10b. Shared routing docs must describe the current 2-skill surface.
+#      A stale auto-resolve/preflight/ideate-CHALLENGE reference can misroute
+#      bounded pair work back into unisolated or retired Codex paths.
+# ---------------------------------------------------------------------------
+section "Check 10b: Shared routing docs avoid retired skill surfaces"
+offenders=$(grep -RInE 'auto-resolve/SKILL\.md|preflight/SKILL\.md|challenge-rubric\.md|ideate CHALLENGE phase|does NOT consume this file|cross-model challenge phases when configured|phase-1-build\.md|phase-2-evaluate\.md|phase-3-critic\.md' \
+  config/skills/_shared 2>/dev/null || true)
+if [ -z "$offenders" ]; then
+  ok "shared routing docs reference current devlyn:ideate/devlyn:resolve surface"
+else
+  while IFS= read -r f; do bad "$f"; done <<< "$offenders"
+fi
+
+# ---------------------------------------------------------------------------
+# 10c. User-facing current docs must not advertise retired skills.
+#      Historical archive mentions are allowed, but install/package copy and
+#      active helper scripts should describe ideate -> resolve, not
+#      ideate -> auto-resolve -> preflight or ideate CHALLENGE.
+# ---------------------------------------------------------------------------
+section "Check 10c: User-facing current docs avoid retired skill surfaces"
+offenders=$(
+  {
+    grep -nE '/devlyn:auto-resolve|ideate CHALLENGE|--with-codex|Quick Start pointing to ideate → auto-resolve → preflight|auto-resolve'\''s build agent|Core pipeline skills \(`ideate`, `auto-resolve`, `preflight`\)' README.md 2>/dev/null || true
+    grep -nE '"description": .*auto-resolve|"description": .*preflight' package.json 2>/dev/null || true
+    grep -nE 'so auto-resolve doesn'\''t prompt' bin/devlyn.js 2>/dev/null || true
+    grep -nE 'devlyn:auto-resolve|phase-1-build\.md|phase-2-evaluate\.md|phase-3-critic\.md' scripts/static-ab.sh 2>/dev/null || true
+  } | sed -E 's#^#user-facing retired-surface reference: #'
+)
+if [ -z "$offenders" ]; then
+  ok "README/package/helper copy describes current ideate -> resolve surface"
+else
+  while IFS= read -r f; do bad "$f"; done <<< "$offenders"
+fi
+
+# ---------------------------------------------------------------------------
+# 10d. Prompt adapters cite the current official model guidance.
+#      Prompt edits must be model-specific where tactics differ: OpenAI guidance
+#      for GPT/Codex, Anthropic guidance for Claude. This check keeps that
+#      contract from becoming oral tradition.
+# ---------------------------------------------------------------------------
+section "Check 10d: Prompt adapters cite official model guidance"
+adapter_missing=0
+if ! grep -Fq 'https://developers.openai.com/api/docs/guides/prompt-guidance?model=gpt-5.5' \
+  config/skills/_shared/adapters/gpt-5-5.md; then
+  bad "gpt-5-5 adapter missing official OpenAI prompt guidance URL"
+  adapter_missing=1
+fi
+if ! grep -Fq 'https://platform.claude.com/docs/en/build-with-claude/prompt-engineering/claude-prompting-best-practices' \
+  config/skills/_shared/adapters/opus-4-7.md; then
+  bad "opus-4-7 adapter missing official Claude prompting best-practices URL"
+  adapter_missing=1
+fi
+for pattern in 'Use Markdown only where it carries structure' 'metaprompter for itself'; do
+  if ! grep -Fq "$pattern" config/skills/_shared/adapters/gpt-5-5.md; then
+    bad "gpt-5-5 adapter missing official-guidance tactic: $pattern"
+    adapter_missing=1
+  fi
+done
+for pattern in 'high` or `xhigh` effort' 'report every issue you find' 'do not filter for importance or confidence' 'prefer concise positive examples' '<example>'; do
+  if ! grep -Fq "$pattern" config/skills/_shared/adapters/opus-4-7.md; then
+    bad "opus-4-7 adapter missing official-guidance tactic: $pattern"
+    adapter_missing=1
+  fi
+done
+for file in config/skills/devlyn:resolve/SKILL.md config/skills/devlyn:ideate/SKILL.md; do
+  if ! grep -Fq '_shared/adapters/<model>.md' "$file"; then
+    bad "$file — missing per-engine adapter injection contract"
+    adapter_missing=1
+  fi
+done
+if [ $adapter_missing -eq 0 ]; then
+  ok "adapters cite official GPT/Claude guidance, carry model-specific tactics, and both skills inject them"
+fi
+
+# ---------------------------------------------------------------------------
+# 10d1. Opus sidecar must fail closed on score-source mapping before provider calls.
+# ---------------------------------------------------------------------------
+section "Check 10d1: Opus judge sidecar validates blind mapping before provider calls"
+if ! grep -Fq 'judge blind mapping missing arm(s)' benchmark/auto-resolve/scripts/judge-opus-pass.sh \
+  || ! grep -Fq 'scores_by_arm without blind mapping' benchmark/auto-resolve/scripts/judge-opus-pass.sh \
+  || ! grep -Fq 'scores_by_arm malformed score(s)' benchmark/auto-resolve/scripts/judge-opus-pass.sh \
+  || ! grep -Fq 'gpt judge.json _blind_mapping must be an object' benchmark/auto-resolve/scripts/judge-opus-pass.sh \
+  || ! grep -Fq 'def is_score(value):' benchmark/auto-resolve/scripts/judge-opus-pass.sh \
+  || ! grep -Fq 'not isinstance(value, bool) and 0 <= value <= 100' benchmark/auto-resolve/scripts/judge-opus-pass.sh \
+  || ! grep -Fq 'invalid opus score value(s)' benchmark/auto-resolve/scripts/judge-opus-pass.sh \
+  || ! grep -Fq 'invalid opus disqualifier value(s)' benchmark/auto-resolve/scripts/judge-opus-pass.sh \
+  || ! grep -Fq 'opus-invalid-generated-dq' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'def blind_mapping(j):' benchmark/auto-resolve/scripts/judge-opus-pass.sh \
+  || ! grep -Fq 'def mapped_arm_set(j):' benchmark/auto-resolve/scripts/judge-opus-pass.sh \
+  || ! grep -Fq '{"solo_claude", "bare"}.issubset(mapped_arms)' benchmark/auto-resolve/scripts/judge-opus-pass.sh \
+  || ! grep -Fq 'def mapped_scores(j):' benchmark/auto-resolve/scripts/judge-opus-pass.sh \
+  || ! grep -Fq 'if arm in mapped_arms and is_score(score)' benchmark/auto-resolve/scripts/judge-opus-pass.sh \
+  || ! grep -Fq 'def margin_from_scores(scores, left, right):' benchmark/auto-resolve/scripts/judge-opus-pass.sh \
+  || ! grep -Fq 'def mapped_winner(j, scores):' benchmark/auto-resolve/scripts/judge-opus-pass.sh \
+  || ! grep -Fq 'def fmt_metric(value):' benchmark/auto-resolve/scripts/judge-opus-pass.sh \
+  || ! grep -Fq 'g_l1_l0 = margin_from_scores(g_scores, "solo_claude", "bare")' benchmark/auto-resolve/scripts/judge-opus-pass.sh \
+  || ! grep -Fq 'g_v_l0 = margin_from_scores(g_scores, "variant", "bare")' benchmark/auto-resolve/scripts/judge-opus-pass.sh \
+  || ! grep -Fq 'g_winner = mapped_winner(g, g_scores)' benchmark/auto-resolve/scripts/judge-opus-pass.sh \
+  || ! grep -Fq '"winner_agree":   g_winner is not None and o_winner is not None and g_winner == o_winner' benchmark/auto-resolve/scripts/judge-opus-pass.sh \
+  || ! grep -Fq '"gpt_scores":   g_scores' benchmark/auto-resolve/scripts/judge-opus-pass.sh \
+  || ! grep -Fq '"opus_scores":  o_scores' benchmark/auto-resolve/scripts/judge-opus-pass.sh \
+  || ! grep -Fq 'opus-bad-mapping' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'opus-malformed-mapping' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'opus-malformed-score' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'opus-invalid-generated-score' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'arg-parse-opus-summary-mapping' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'arg-parse-opus-summary-null-margin' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'assert row["gpt_margin_l1_l0"] == 10' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'assert row["gpt_winner"] is None' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq "gpt_l1_l0_avg=na" benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'solo_claude-bare={chosen' benchmark/auto-resolve/scripts/judge-opus-pass.sh \
+  || ! grep -Fq 'variant-solo_claude={chosen' benchmark/auto-resolve/scripts/judge-opus-pass.sh; then
+  offenders="${offenders}"$'\n'"judge-opus-pass.sh must validate _blind_mapping locally before invoking Claude"
+fi
+
+# ---------------------------------------------------------------------------
+# 10e. Benchmark docs must describe the current solo-vs-pair topology.
+#      Pair evidence work depends on the bare/solo_claude/pair contract being current:
+#      bare, solo_claude, and the selected pair arm are measured today.
+# ---------------------------------------------------------------------------
+section "Check 10e: Benchmark docs describe current 3-arm pair topology"
+offenders=$(grep -RInE 'L1 .*queued|cannot directly verify the L1|auto-resolve → preflight|ideate → auto-resolve|all fixtures × 2 arms|9 fixtures × 2 arms|9 fixtures × 3 arms|≥ 7 of 9|7/9 fixtures|variant` / `bare|variant/\{input|preflight; bare|Audited by preflight|Variant'\''s CRITIC|future enhancement|release-blocker|Today the suite runs `variant`|/devlyn:auto-resolve|auto-resolve-ready|REAL auto-resolve|real auto-resolve|auto-resolve run|auto-resolve runs|Claude/auto-resolve|variant − bare|Margin \(variant|Ship thresholds use margin \(variant|both arms improve together|both arms|benchmark --n 3|run-suite\.sh --n 3|higher confidence for ship decisions|3 runs per fixture for ship decisions|One-command A/B benchmark|A/B randomized|A/B benchmark suite vs bare|not vs bare — bare is the opponent' \
+  benchmark/auto-resolve/BENCHMARK-DESIGN.md \
+  benchmark/auto-resolve/README.md \
+  benchmark/auto-resolve/RUBRIC.md \
+  benchmark/auto-resolve/BENCHMARK-RESULTS.md \
+  benchmark/auto-resolve/run-real-benchmark.md \
+  benchmark/auto-resolve/fixtures/SCHEMA.md \
+  benchmark/auto-resolve/shadow-fixtures/README.md \
+	  2>/dev/null || true)
+fixture_note_stale=$(git grep -InE -- 'both arms|both solo and pair arms|bare or solo consistently reaches ceiling' -- \
+  'benchmark/auto-resolve/fixtures/**/NOTES.md' \
+  ':!benchmark/auto-resolve/fixtures/retired/**' \
+  2>/dev/null || true)
+if [ -n "$fixture_note_stale" ]; then
+  offenders="${offenders}"$'\n'"fixture NOTES must name bare and solo_claude explicitly instead of ambiguous arm wording:"$'\n'"$fixture_note_stale"
+fi
+shadow_note_stale=$(git grep -In -- 'solo headroom' -- \
+  'benchmark/auto-resolve/shadow-fixtures/**/NOTES.md' \
+  2>/dev/null || true)
+if [ -n "$shadow_note_stale" ]; then
+  offenders="${offenders}"$'\n'"shadow fixture NOTES must name solo_claude headroom explicitly:"$'\n'"$shadow_note_stale"
+fi
+active_doc_stale_solo_scores=$(git grep -InE -- '(^|[^[:alnum:]_])solo [0-9]+|/ solo$|, solo [0-9]+|vs solo [0-9]+' -- \
+  benchmark/auto-resolve/BENCHMARK-RESULTS.md \
+  benchmark/auto-resolve/README.md \
+  benchmark/auto-resolve/run-real-benchmark.md \
+  2>/dev/null || true)
+if [ -n "$active_doc_stale_solo_scores" ]; then
+  offenders="${offenders}"$'\n'"active benchmark score evidence must name solo_claude instead of shorthand solo:"$'\n'"$active_doc_stale_solo_scores"
+fi
+active_stale_margin_labels=$(git grep -InE -- 'l1-l0=|v-l1=|L1-L0 disagreement|Per-axis L1-L0|Suite-level per-axis L1-L0|Suite avg L1-L0|L2-L1 margin' -- \
+  benchmark/auto-resolve/BENCHMARK-DESIGN.md \
+  benchmark/auto-resolve/README.md \
+  benchmark/auto-resolve/RUBRIC.md \
+  benchmark/auto-resolve/scripts/judge-opus-pass.sh \
+  2>/dev/null || true)
+if [ -n "$active_stale_margin_labels" ]; then
+  offenders="${offenders}"$'\n'"active benchmark docs/stdout must use key-aligned margin labels:"$'\n'"$active_stale_margin_labels"
+fi
+if grep -RInE 'gate: ≥ 7 of 9|Hard floor 3: ≥ 7 of 9|7-of-9 L1 floor|all 9 fixtures produced' \
+  benchmark/auto-resolve/scripts/compile-report.py \
+  benchmark/auto-resolve/scripts/ship-gate.py \
+  benchmark/auto-resolve/scripts/judge-opus-pass.sh >/dev/null 2>&1; then
+  offenders="${offenders}"$'\n'"benchmark scripts must describe the current extended fixture count as an explicit selected/gated set, not stale 9-fixture wording"
+fi
+if grep -Fq 'PAIR_ARM="l2_gated"' benchmark/auto-resolve/scripts/run-full-pipeline-pair-candidate.sh; then
+  offenders="${offenders}"$'\n'"benchmark/auto-resolve/scripts/run-full-pipeline-pair-candidate.sh: default pair arm must stay on current measured l2_risk_probes path"
+fi
+if ! grep -Fq 'for required in result.json verify.json diff.patch; do' \
+  benchmark/auto-resolve/scripts/run-full-pipeline-pair-candidate.sh \
+  || ! grep -Fq 'reuse source missing $required' \
+    benchmark/auto-resolve/scripts/run-full-pipeline-pair-candidate.sh; then
+  offenders="${offenders}"$'\n'"benchmark/auto-resolve/scripts/run-full-pipeline-pair-candidate.sh: calibrated-arm reuse must fail closed on missing result.json, verify.json, and diff.patch"
+fi
+if ! grep -Fq 'reuse destination incomplete $required' \
+  benchmark/auto-resolve/scripts/run-full-pipeline-pair-candidate.sh \
+  || ! grep -Fq 'reuse destination is not a directory' \
+    benchmark/auto-resolve/scripts/run-full-pipeline-pair-candidate.sh; then
+  offenders="${offenders}"$'\n'"benchmark/auto-resolve/scripts/run-full-pipeline-pair-candidate.sh: calibrated-arm reuse must fail closed on incomplete existing destination"
+fi
+if ! grep -Fq 'diff.patch missing' benchmark/auto-resolve/scripts/headroom-gate.py \
+  || ! grep -Fq 'diff.patch missing' benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py; then
+  offenders="${offenders}"$'\n'"benchmark pair gates must require diff.patch artifacts for measured arms"
+fi
+if ! grep -Fq 'pair-arm must be l2_risk_probes or l2_gated' \
+  benchmark/auto-resolve/scripts/run-full-pipeline-pair-candidate.sh \
+  || ! grep -Fq 'pair-arm l2_forced is retired' \
+    benchmark/auto-resolve/scripts/run-full-pipeline-pair-candidate.sh; then
+  offenders="${offenders}"$'\n'"benchmark/auto-resolve/scripts/run-full-pipeline-pair-candidate.sh: pair arm selection must fail closed before fixture execution"
+fi
+if ! grep -Fq 'l2_risk_probes|l2_gated) ;;' \
+  benchmark/auto-resolve/scripts/run-full-pipeline-pair-candidate.sh; then
+  offenders="${offenders}"$'\n'"benchmark/auto-resolve/scripts/run-full-pipeline-pair-candidate.sh: runner allowlist must remain l2_risk_probes|l2_gated only"
+fi
+if grep -Fq 'default="l2_gated"' benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py; then
+  offenders="${offenders}"$'\n'"benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py: --pair-arm default must stay on current measured l2_risk_probes path"
+fi
+if ! grep -Fq 'from pair_evidence_contract import (' \
+  benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py \
+  || ! grep -Fq 'ALLOWED_PAIR_ARMS,' benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py \
+  || ! grep -Fq 'ALLOWED_PAIR_ARMS = {"l2_risk_probes", "l2_gated"}' \
+    benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'pair-arm l2_forced is retired' \
+    benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py; then
+  offenders="${offenders}"$'\n'"benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py: pair arm selection must fail closed inside the gate"
+fi
+if ! grep -Fq '"l2_risk_probes"' benchmark/auto-resolve/scripts/check-f9-artifacts.py; then
+  offenders="${offenders}"$'\n'"benchmark/auto-resolve/scripts/check-f9-artifacts.py: F9 skill-driven artifact check must accept current l2_risk_probes arm"
+fi
+if ! grep -Fq '_load_json_object' benchmark/auto-resolve/scripts/check-f9-artifacts.py \
+  || ! grep -Fq 'expected JSON object' benchmark/auto-resolve/scripts/test-check-f9-artifacts.sh; then
+  offenders="${offenders}"$'\n'"benchmark/auto-resolve/scripts/check-f9-artifacts.py: F9 timing/state JSON must fail closed on non-object payloads"
+fi
+if grep -RInE 'asserts variant/solo|Variant-only artifact checks|Variant artifact check' \
+  benchmark/auto-resolve/scripts/check-f9-artifacts.py \
+  benchmark/auto-resolve/fixtures/F9-e2e-ideate-to-resolve/NOTES.md \
+  benchmark/auto-resolve/fixtures/F9-e2e-ideate-to-resolve/spec.md >/dev/null 2>&1; then
+  offenders="${offenders}"$'\n'"F9 artifact docs/checker wording must describe skill-driven arms, not variant-only checks"
+fi
+if grep -Fq '<variant|bare>' benchmark/auto-resolve/scripts/run-fixture.sh; then
+  offenders="${offenders}"$'\n'"benchmark/auto-resolve/scripts/run-fixture.sh: usage must list current benchmark arms"
+fi
+if grep -Fq 'l2_gated/l2_forced' benchmark/auto-resolve/scripts/run-fixture.sh; then
+  offenders="${offenders}"$'\n'"benchmark/auto-resolve/scripts/run-fixture.sh: comments must not omit l2_risk_probes from l2_* arm handling"
+fi
+if grep -Fq 'ENGINE_CLAUSE="--engine auto"' benchmark/auto-resolve/scripts/run-fixture.sh \
+  || grep -Fq 'Run with `--engine auto`' benchmark/auto-resolve/scripts/run-fixture.sh; then
+  offenders="${offenders}"$'\n'"benchmark/auto-resolve/scripts/run-fixture.sh: smoke variant arm must use current --engine claude --risk-probes path, not retired --engine auto"
+fi
+if ! grep -Fq 'ENGINE_CLAUSE="--engine claude --risk-probes"' benchmark/auto-resolve/scripts/run-fixture.sh; then
+  offenders="${offenders}"$'\n'"benchmark/auto-resolve/scripts/run-fixture.sh: variant/l2 risk-probes path must remain available"
+fi
+if ! grep -Fq '{bare, solo_claude, selected pair arm}' benchmark/auto-resolve/scripts/judge.sh; then
+  offenders="${offenders}"$'\n'"benchmark/auto-resolve/scripts/judge.sh: blind judge topology comment must describe current pair-candidate proof runs"
+fi
+if grep -Fq 'blind judge scores l2_gated' benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py \
+  || grep -Fq 'l2_gated is clean' benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py; then
+  offenders="${offenders}"$'\n'"benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py: gate docstring must describe selected pair arm, not l2_gated-only proof"
+fi
+if grep -Fq 'only then spends a `l2_gated` arm' benchmark/auto-resolve/README.md \
+  || grep -Fq 'fresh `l2_gated` measurement' benchmark/auto-resolve/README.md; then
+  offenders="${offenders}"$'\n'"benchmark/auto-resolve/README.md: full-pipeline runner docs must not present l2_gated as the default measured proof path"
+fi
+if ! grep -Fq 'Pair arms are limited to current' benchmark/auto-resolve/README.md \
+  || ! grep -Fq '`l2_forced` is' benchmark/auto-resolve/README.md; then
+  offenders="${offenders}"$'\n'"benchmark/auto-resolve/README.md: full-pipeline docs must state allowed pair arms and retired l2_forced"
+fi
+if ! grep -Fq 'Current solo<pair' benchmark/auto-resolve/RUBRIC.md \
+  || ! grep -Fq 'evidence uses the full-pipeline pair gate' benchmark/auto-resolve/RUBRIC.md \
+  || ! grep -Fq 'explicit selected pair arm' benchmark/auto-resolve/RUBRIC.md \
+  || ! grep -Fq 'selected pair arm over `solo_claude`' benchmark/auto-resolve/RUBRIC.md; then
+  offenders="${offenders}"$'\n'"benchmark/auto-resolve/RUBRIC.md: rubric must distinguish legacy run-suite ship gate from current selected-pair-arm evidence"
+fi
+if grep -Fq 'PHASE 8' benchmark/auto-resolve/run-real-benchmark.md \
+  || grep -Fq 'security_review on `--engine auto`' benchmark/auto-resolve/run-real-benchmark.md \
+  || grep -Fq 'BENCHMARK-RESULTS-v3.md' benchmark/auto-resolve/run-real-benchmark.md; then
+  offenders="${offenders}"$'\n'"benchmark/auto-resolve/run-real-benchmark.md: real-run docs must describe the current 5-phase resolve and pair score harness"
+fi
+if ! grep -Fq 'Archive note (2026-05-14): historical pre-cutover benchmark plan' benchmark/auto-resolve/v3.6-ab-plan.md \
+  || ! grep -Fq 'Do not use this file for current' benchmark/auto-resolve/v3.6-ab-plan.md \
+  || ! grep -Fq 'Archive note (2026-05-14): historical pre-cutover results' benchmark/auto-resolve/BENCHMARK-RESULTS-v3.md \
+  || ! grep -Fq 'Do not treat these projected' benchmark/auto-resolve/BENCHMARK-RESULTS-v3.md \
+  || ! grep -Fq 'Archive note (2026-05-14): historical v3.2 pilot' benchmark/auto-resolve/PILOT-RESULTS-v3.2.md \
+  || ! grep -Fq 'This n=1 pilot is not current solo<pair evidence' benchmark/auto-resolve/PILOT-RESULTS-v3.2.md \
+  || ! grep -Fq 'Archive note (2026-05-14): historical v3.2 strict-route pilot' benchmark/auto-resolve/PILOT-RESULTS-STRICT-v3.2.md \
+  || ! grep -Fq 'This inline n=1 pilot is not current solo<pair evidence' benchmark/auto-resolve/PILOT-RESULTS-STRICT-v3.2.md \
+  || ! grep -Fq 'archived static comparison helper for pre-cutover auto-resolve' benchmark/auto-resolve/measure-static.py \
+  || ! grep -Fq 'is not current solo<pair evidence' benchmark/auto-resolve/measure-static.py; then
+  offenders="${offenders}"$'\n'"archived v3 benchmark artifacts must be clearly marked as non-current solo<pair evidence"
+fi
+if ! grep -Fq 'l2_risk_probes` | current measured pair path' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'Dry-runs, lint, and shell tests prove wiring only' benchmark/auto-resolve/run-real-benchmark.md; then
+  offenders="${offenders}"$'\n'"benchmark/auto-resolve/run-real-benchmark.md: real-run docs must name the measured pair arm and distinguish wiring checks from scores"
+fi
+if grep -RInE 'F27.*unmeasured|unmeasured.*F27' \
+  benchmark/auto-resolve/README.md \
+  benchmark/auto-resolve/BENCHMARK-RESULTS.md \
+  benchmark/auto-resolve/run-real-benchmark.md \
+  benchmark/auto-resolve/fixtures/retired/F27-cli-subscription-proration/NOTES.md >/dev/null 2>&1; then
+  offenders="${offenders}"$'\n'"F27 docs must record the measured headroom failure, not stale unmeasured-candidate wording"
+fi
+if [ -d benchmark/auto-resolve/fixtures/F27-cli-subscription-proration ]; then
+  offenders="${offenders}"$'\n'"F27 must stay out of active golden fixtures after failing headroom; keep it under fixtures/retired/"
+fi
+if [ ! -f benchmark/auto-resolve/fixtures/retired/F27-cli-subscription-proration/RETIRED.md ]; then
+  offenders="${offenders}"$'\n'"retired F27 must keep RETIRED.md with the measured rejection reason"
+fi
+if ! grep -Fq '"benchmark/auto-resolve/fixtures/retired/F*/**"' package.json; then
+  offenders="${offenders}"$'\n'"package.json must include retired fixtures so replay artifacts ship in npm packages"
+fi
+if ! grep -Fq '20260511-f27-headroom-smoke-061401' benchmark/auto-resolve/fixtures/retired/F27-cli-subscription-proration/NOTES.md \
+  || ! grep -Fq '20260511-f27-headroom-smoke-061401' benchmark/auto-resolve/fixtures/retired/F27-cli-subscription-proration/RETIRED.md; then
+  offenders="${offenders}"$'\n'"F27 notes must cite the measured headroom smoke run"
+fi
+if grep -Fq 'better candidate for pair risk probes' benchmark/auto-resolve/fixtures/retired/F27-cli-subscription-proration/NOTES.md; then
+  offenders="${offenders}"$'\n'"F27 notes must not describe rejected F27 as a better pair-risk-probe candidate"
+fi
+if grep -Fq 'F16-cli-quote-tax-rules F27-cli-subscription-proration' \
+  benchmark/auto-resolve/README.md benchmark/auto-resolve/run-real-benchmark.md; then
+  offenders="${offenders}"$'\n'"benchmark docs must not recommend rejected F27 in headroom/full-pipeline command examples"
+fi
+if [ -d benchmark/auto-resolve/fixtures/F28-cli-return-authorization ]; then
+  offenders="${offenders}"$'\n'"F28 must stay out of active golden fixtures after corrected-oracle reverify failed headroom; keep it under fixtures/retired/"
+fi
+if [ ! -f benchmark/auto-resolve/fixtures/retired/F28-cli-return-authorization/RETIRED.md ]; then
+  offenders="${offenders}"$'\n'"retired F28 must keep RETIRED.md with the corrected-oracle rejection reason"
+fi
+if ! grep -Fq '20260511-f28-headroom-smoke-085307' benchmark/auto-resolve/fixtures/retired/F28-cli-return-authorization/NOTES.md \
+  || ! grep -Fq '20260511-f28-pair-smoke-091021' benchmark/auto-resolve/fixtures/retired/F28-cli-return-authorization/NOTES.md \
+  || ! grep -Fq '20260511-f28-policy-oraclefix-reverified-pair' benchmark/auto-resolve/fixtures/retired/F28-cli-return-authorization/NOTES.md \
+  || ! grep -Fq '20260511-f28-policy-oraclefix-reverified-pair' benchmark/auto-resolve/fixtures/retired/F28-cli-return-authorization/RETIRED.md \
+  || ! grep -Fq '20260511-f28-headroom-smoke-085307' benchmark/auto-resolve/BENCHMARK-RESULTS.md \
+  || ! grep -Fq '20260511-f28-pair-smoke-091021' benchmark/auto-resolve/BENCHMARK-RESULTS.md \
+  || ! grep -Fq '20260511-f28-headroom-smoke-085307' benchmark/auto-resolve/README.md \
+  || ! grep -Fq '20260511-f28-pair-smoke-091021' benchmark/auto-resolve/README.md \
+  || ! grep -Fq '20260511-f28-headroom-smoke-085307' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq '20260511-f28-pair-smoke-091021' benchmark/auto-resolve/run-real-benchmark.md; then
+  offenders="${offenders}"$'\n'"F28 docs must cite measured smoke and corrected-oracle rejection run ids before anyone counts it"
+fi
+if grep -Fq 'F16-cli-quote-tax-rules F28-cli-return-authorization' \
+  benchmark/auto-resolve/README.md benchmark/auto-resolve/run-real-benchmark.md README.md; then
+  offenders="${offenders}"$'\n'"benchmark docs must not recommend unstable F28 in pair-evidence command examples"
+fi
+if ! bash -n benchmark/auto-resolve/scripts/pair-rejected-fixtures.sh >/dev/null 2>&1; then
+  offenders="${offenders}"$'\n'"pair-rejected-fixtures.sh must be valid bash"
+fi
+if ! grep -Fq 'trivial calibration fixture' benchmark/auto-resolve/scripts/pair-rejected-fixtures.sh \
+  || ! grep -Fq '20260512-f2-medium-headroom' benchmark/auto-resolve/scripts/pair-rejected-fixtures.sh \
+  || ! grep -Fq '20260511-f3-http-error-headroom' benchmark/auto-resolve/scripts/pair-rejected-fixtures.sh \
+  || ! grep -Fq '20260512-f4-web-headroom' benchmark/auto-resolve/scripts/pair-rejected-fixtures.sh \
+  || ! grep -Fq '20260512-f5-fixloop-headroom' benchmark/auto-resolve/scripts/pair-rejected-fixtures.sh \
+  || ! grep -Fq '20260512-f6-checksum-headroom' benchmark/auto-resolve/scripts/pair-rejected-fixtures.sh \
+  || ! grep -Fq '20260512-f7-scope-headroom' benchmark/auto-resolve/scripts/pair-rejected-fixtures.sh \
+  || ! grep -Fq 'known-limit ambiguity fixture' benchmark/auto-resolve/scripts/pair-rejected-fixtures.sh \
+  || ! grep -Fq '20260512-f9-e2e-headroom' benchmark/auto-resolve/scripts/pair-rejected-fixtures.sh \
+  || ! grep -Fq '20260507-f10-f11-tier1-full-pipeline' benchmark/auto-resolve/scripts/pair-rejected-fixtures.sh \
+  || ! grep -Fq '20260511-f12-webhook-headroom' benchmark/auto-resolve/scripts/pair-rejected-fixtures.sh \
+  || ! grep -Fq '20260511-f15-concurrency-headroom' benchmark/auto-resolve/scripts/pair-rejected-fixtures.sh \
+  || ! grep -Fq 'bare 94 / solo_claude 98 in 20260508-f22-exact-error-headroom' benchmark/auto-resolve/scripts/pair-rejected-fixtures.sh \
+  || ! grep -Fq '20260508-f26-headroom' benchmark/auto-resolve/scripts/pair-rejected-fixtures.sh \
+  || ! grep -Fq '20260511-f27-headroom-smoke-061401' benchmark/auto-resolve/scripts/pair-rejected-fixtures.sh \
+  || ! grep -Fq '20260511-f28-policy-oraclefix-reverified-pair' benchmark/auto-resolve/scripts/pair-rejected-fixtures.sh \
+  || ! grep -Fq '20260510-f29-headroom-v2' benchmark/auto-resolve/scripts/pair-rejected-fixtures.sh \
+  || ! grep -Fq '20260511-f30-headroom-v1' benchmark/auto-resolve/scripts/pair-rejected-fixtures.sh \
+  || ! grep -Fq '20260512-f31-seat-rebalance-headroom' benchmark/auto-resolve/scripts/pair-rejected-fixtures.sh \
+  || ! grep -Fq '20260512-f32-subscription-renewal-headroom' benchmark/auto-resolve/scripts/pair-rejected-fixtures.sh \
+  || ! grep -Fq '20260513-s2-inventory-headroom' benchmark/auto-resolve/scripts/pair-rejected-fixtures.sh \
+  || ! grep -Fq '20260513-s3-ticket-headroom' benchmark/auto-resolve/scripts/pair-rejected-fixtures.sh \
+  || ! grep -Fq '20260513-s4-return-headroom' benchmark/auto-resolve/scripts/pair-rejected-fixtures.sh \
+  || ! grep -Fq '20260513-s5-credit-headroom' benchmark/auto-resolve/scripts/pair-rejected-fixtures.sh \
+  || ! grep -Fq '20260514-s6-refund-headroom-v1' benchmark/auto-resolve/scripts/pair-rejected-fixtures.sh \
+  || ! grep -Fq 'source "$BENCH_ROOT/scripts/pair-rejected-fixtures.sh"' benchmark/auto-resolve/scripts/run-headroom-candidate.sh \
+  || ! grep -Fq 'source "$BENCH_ROOT/scripts/pair-rejected-fixtures.sh"' benchmark/auto-resolve/scripts/run-full-pipeline-pair-candidate.sh \
+  || ! grep -Fq 'fixture_smoke_only' benchmark/auto-resolve/scripts/run-headroom-candidate.sh \
+  || ! grep -Fq 'fixture_smoke_only' benchmark/auto-resolve/scripts/run-full-pipeline-pair-candidate.sh \
+  || ! grep -Fq 'smoke-only-s1-provider-run' benchmark/auto-resolve/scripts/test-run-headroom-candidate.sh \
+  || ! grep -Fq 'smoke-only-s1-provider-run' benchmark/auto-resolve/scripts/test-run-full-pipeline-pair-candidate.sh \
+  || ! grep -Fq 'smoke-only-s1-cli-headroom' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'smoke-only-s1-cli-pair' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'declare -F rejected_pair_fixture_reason' benchmark/auto-resolve/scripts/run-headroom-candidate.sh \
+  || ! grep -Fq 'declare -F rejected_pair_fixture_reason' benchmark/auto-resolve/scripts/run-full-pipeline-pair-candidate.sh \
+  || ! grep -Fq 'rejected-f31-fixture' benchmark/auto-resolve/scripts/test-run-headroom-candidate.sh \
+  || ! grep -Fq 'rejected-f32-fixture' benchmark/auto-resolve/scripts/test-run-headroom-candidate.sh \
+  || ! grep -Fq 'rejected-s6-shadow-fixture' benchmark/auto-resolve/scripts/test-run-headroom-candidate.sh \
+  || ! grep -Fq 'rejected-f31-fixture' benchmark/auto-resolve/scripts/test-run-full-pipeline-pair-candidate.sh \
+  || ! grep -Fq 'rejected-f32-fixture' benchmark/auto-resolve/scripts/test-run-full-pipeline-pair-candidate.sh \
+  || ! grep -Fq 'rejected-s6-shadow-fixture' benchmark/auto-resolve/scripts/test-run-full-pipeline-pair-candidate.sh \
+  || ! grep -Fq 'load_rejected_short_ids' benchmark/auto-resolve/scripts/build-pair-eligible-manifest.py \
+  || ! grep -Fq 'load_rejected_fixture_reasons' benchmark/auto-resolve/scripts/build-pair-eligible-manifest.py \
+  || ! grep -Fq 'rejected_excluded' benchmark/auto-resolve/scripts/build-pair-eligible-manifest.py \
+  || ! grep -Fq 'rejected_excluded_reasons' benchmark/auto-resolve/scripts/build-pair-eligible-manifest.py \
+  || ! grep -Fq 'rejected_excluded_reasons' benchmark/auto-resolve/scripts/test-build-pair-eligible-manifest.sh \
+  || ! grep -Fq 'selection_rule.rejected_excluded_reasons' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'rejected_excluded_reasons' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'fixtures_pair_eligible must not be empty' benchmark/auto-resolve/scripts/iter-0033c-compare.py \
+  || ! grep -Fq 'gate3_threshold_count must be a positive integer' benchmark/auto-resolve/scripts/iter-0033c-compare.py \
+  || ! grep -Fq 'selection_rule.rejected_excluded_reasons keys must match rejected_excluded' benchmark/auto-resolve/scripts/iter-0033c-compare.py \
+  || ! grep -Fq 'nan-threshold-manifest' benchmark/auto-resolve/scripts/test-iter-0033c-compare.sh \
+  || ! grep -Fq 'bad-rejected-reasons-manifest' benchmark/auto-resolve/scripts/test-iter-0033c-compare.sh \
+  || ! grep -Fq 'fixture rejected for pair-candidate runs' benchmark/auto-resolve/scripts/headroom-gate.py \
+  || ! grep -Fq 'fixture rejected for pair-candidate runs' benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py \
+  || ! grep -Fq 'rejected fixture registry missing' benchmark/auto-resolve/scripts/headroom-gate.py \
+  || ! grep -Fq 'rejected fixture registry missing' benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py \
+  || ! grep -Fq '([FS]\d+)-\*\|([FS]\d+)' benchmark/auto-resolve/scripts/headroom-gate.py \
+  || ! grep -Fq '([FS]\d+)-\*\|([FS]\d+)' benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py \
+  || ! grep -Fq '([FS]\d+)-\*\|([FS]\d+)' benchmark/auto-resolve/scripts/pair-candidate-frontier.py \
+  || ! grep -Fq '([FS]\d+)-\*\|([FS]\d+)' benchmark/auto-resolve/scripts/audit-headroom-rejections.py \
+  || ! grep -Fq '([FS]\d+)-\*\|([FS]\d+)' benchmark/auto-resolve/scripts/build-pair-eligible-manifest.py \
+  || ! grep -Fq 'missing-rejected-registry' benchmark/auto-resolve/scripts/test-headroom-gate.sh \
+  || ! grep -Fq 'missing-rejected-registry' benchmark/auto-resolve/scripts/test-full-pipeline-pair-gate.sh \
+  || ! grep -Fq 'rejected-shadow-direct' benchmark/auto-resolve/scripts/test-headroom-gate.sh \
+  || ! grep -Fq 'rejected-shadow-direct' benchmark/auto-resolve/scripts/test-full-pipeline-pair-gate.sh \
+  || ! grep -Fq 's-only-registry' benchmark/auto-resolve/scripts/test-pair-candidate-frontier.sh \
+  || ! grep -Fq 's-only-registry' benchmark/auto-resolve/scripts/test-audit-headroom-rejections.sh \
+  || ! grep -Fq 'shadow solo ceiling' benchmark/auto-resolve/scripts/test-build-pair-eligible-manifest.sh \
+  || grep -Fq 'case "$fid" in' benchmark/auto-resolve/scripts/run-headroom-candidate.sh \
+  || grep -Fq 'case "$fid" in' benchmark/auto-resolve/scripts/run-full-pipeline-pair-candidate.sh; then
+  offenders="${offenders}"$'\n'"pair candidate runners, audits, and manifest builder must honor the shared rejected fixture registry, including F* fixtures and S* shadow controls, without duplicating the case table"
+fi
+if ! grep -Fq 'Pair-candidate status: rejected by design' benchmark/auto-resolve/fixtures/F1-cli-trivial-flag/NOTES.md \
+  || ! grep -Fq 'Pair-candidate status: rejected by design' benchmark/auto-resolve/fixtures/F8-known-limit-ambiguous/NOTES.md \
+  || ! grep -Fq 'rejected by design' benchmark/auto-resolve/BENCHMARK-RESULTS.md \
+  || ! grep -Fq 'calibration/known-limit controls' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'calibration/known-limit controls' benchmark/auto-resolve/run-real-benchmark.md; then
+  offenders="${offenders}"$'\n'"F1/F8 docs must mark calibration and known-limit controls as rejected by design for pair evidence"
+fi
+if ! grep -Fq '20260512-f2-medium-headroom' benchmark/auto-resolve/fixtures/F2-cli-medium-subcommand/NOTES.md \
+  || ! grep -Fq '20260512-f2-medium-headroom' benchmark/auto-resolve/BENCHMARK-RESULTS.md \
+  || ! grep -Fq '20260512-f2-medium-headroom' benchmark/auto-resolve/README.md \
+  || ! grep -Fq '20260512-f2-medium-headroom' benchmark/auto-resolve/run-real-benchmark.md; then
+  offenders="${offenders}"$'\n'"F2 docs must cite the measured headroom rejection before anyone counts it"
+fi
+if ! grep -Fq '20260511-f12-webhook-headroom' benchmark/auto-resolve/fixtures/F12-webhook-raw-body-signature/NOTES.md \
+  || ! grep -Fq '20260511-f12-webhook-headroom' benchmark/auto-resolve/BENCHMARK-RESULTS.md \
+  || ! grep -Fq '20260511-f12-webhook-headroom' benchmark/auto-resolve/README.md \
+  || ! grep -Fq '20260511-f12-webhook-headroom' benchmark/auto-resolve/run-real-benchmark.md; then
+  offenders="${offenders}"$'\n'"F12 docs must cite the measured headroom rejection before anyone counts it"
+fi
+if ! grep -Fq '20260507-f10-f11-tier1-full-pipeline' benchmark/auto-resolve/fixtures/F10-persist-write-collision/NOTES.md \
+  || ! grep -Fq '20260507-f10-f11-tier1-full-pipeline' benchmark/auto-resolve/fixtures/F11-batch-import-all-or-nothing/NOTES.md \
+  || ! grep -Fq '20260507-f10-f11-tier1-full-pipeline' benchmark/auto-resolve/BENCHMARK-RESULTS.md \
+  || ! grep -Fq '20260507-f10-f11-tier1-full-pipeline' benchmark/auto-resolve/README.md \
+  || ! grep -Fq '20260507-f10-f11-tier1-full-pipeline' benchmark/auto-resolve/run-real-benchmark.md; then
+  offenders="${offenders}"$'\n'"F10/F11 docs must cite the measured headroom rejection before anyone counts them"
+fi
+if ! grep -Fq '20260511-f3-http-error-headroom' benchmark/auto-resolve/fixtures/F3-backend-contract-risk/NOTES.md \
+  || ! grep -Fq '20260511-f3-http-error-headroom' benchmark/auto-resolve/BENCHMARK-RESULTS.md \
+  || ! grep -Fq '20260511-f3-http-error-headroom' benchmark/auto-resolve/README.md \
+  || ! grep -Fq '20260511-f3-http-error-headroom' benchmark/auto-resolve/run-real-benchmark.md; then
+  offenders="${offenders}"$'\n'"F3 docs must cite the measured headroom rejection before anyone counts it"
+fi
+if ! grep -Fq '20260512-f4-web-headroom' benchmark/auto-resolve/fixtures/F4-web-browser-design/NOTES.md \
+  || ! grep -Fq '20260512-f4-web-headroom' benchmark/auto-resolve/BENCHMARK-RESULTS.md \
+  || ! grep -Fq '20260512-f4-web-headroom' benchmark/auto-resolve/README.md \
+  || ! grep -Fq '20260512-f4-web-headroom' benchmark/auto-resolve/run-real-benchmark.md; then
+  offenders="${offenders}"$'\n'"F4 docs must cite the measured headroom rejection before anyone counts it"
+fi
+if ! grep -Fq '20260512-f5-fixloop-headroom' benchmark/auto-resolve/fixtures/F5-fix-loop-red-green/NOTES.md \
+  || ! grep -Fq '20260512-f5-fixloop-headroom' benchmark/auto-resolve/BENCHMARK-RESULTS.md \
+  || ! grep -Fq '20260512-f5-fixloop-headroom' benchmark/auto-resolve/README.md \
+  || ! grep -Fq '20260512-f5-fixloop-headroom' benchmark/auto-resolve/run-real-benchmark.md; then
+  offenders="${offenders}"$'\n'"F5 docs must cite the measured headroom rejection before anyone counts it"
+fi
+if ! grep -Fq '20260512-f6-checksum-headroom' benchmark/auto-resolve/fixtures/F6-dep-audit-native-module/NOTES.md \
+  || ! grep -Fq '20260512-f6-checksum-headroom' benchmark/auto-resolve/BENCHMARK-RESULTS.md \
+  || ! grep -Fq '20260512-f6-checksum-headroom' benchmark/auto-resolve/README.md \
+  || ! grep -Fq '20260512-f6-checksum-headroom' benchmark/auto-resolve/run-real-benchmark.md; then
+  offenders="${offenders}"$'\n'"F6 docs must cite the measured headroom rejection before anyone counts it"
+fi
+if ! grep -Fq '20260512-f7-scope-headroom' benchmark/auto-resolve/fixtures/F7-out-of-scope-trap/NOTES.md \
+  || ! grep -Fq '20260512-f7-scope-headroom' benchmark/auto-resolve/BENCHMARK-RESULTS.md \
+  || ! grep -Fq '20260512-f7-scope-headroom' benchmark/auto-resolve/README.md \
+  || ! grep -Fq '20260512-f7-scope-headroom' benchmark/auto-resolve/run-real-benchmark.md; then
+  offenders="${offenders}"$'\n'"F7 docs must cite the measured headroom rejection before anyone counts it"
+fi
+if ! grep -Fq '20260512-f9-e2e-headroom' benchmark/auto-resolve/fixtures/F9-e2e-ideate-to-resolve/NOTES.md \
+  || ! grep -Fq '20260512-f9-e2e-headroom' benchmark/auto-resolve/BENCHMARK-RESULTS.md \
+  || ! grep -Fq '20260512-f9-e2e-headroom' benchmark/auto-resolve/README.md \
+  || ! grep -Fq '20260512-f9-e2e-headroom' benchmark/auto-resolve/run-real-benchmark.md; then
+  offenders="${offenders}"$'\n'"F9 docs must cite the measured headroom rejection before anyone counts it"
+fi
+if ! grep -Fq '20260511-f15-concurrency-headroom' benchmark/auto-resolve/fixtures/F15-frozen-diff-race-review/NOTES.md \
+  || ! grep -Fq '20260511-f15-concurrency-headroom' benchmark/auto-resolve/BENCHMARK-RESULTS.md \
+  || ! grep -Fq '20260511-f15-concurrency-headroom' benchmark/auto-resolve/README.md \
+  || ! grep -Fq '20260511-f15-concurrency-headroom' benchmark/auto-resolve/run-real-benchmark.md; then
+  offenders="${offenders}"$'\n'"F15 docs must cite the measured headroom rejection before anyone counts it"
+fi
+if ! grep -Fq '20260512-f31-seat-rebalance-headroom' benchmark/auto-resolve/fixtures/F31-cli-seat-rebalance/NOTES.md \
+  || ! grep -Fq '20260512-f31-seat-rebalance-headroom' benchmark/auto-resolve/BENCHMARK-RESULTS.md \
+  || ! grep -Fq '20260512-f31-seat-rebalance-headroom' benchmark/auto-resolve/README.md \
+  || ! grep -Fq '20260512-f31-seat-rebalance-headroom' benchmark/auto-resolve/run-real-benchmark.md; then
+  offenders="${offenders}"$'\n'"F31 docs must cite the measured headroom rejection before anyone counts it"
+fi
+if grep -Fq 'execFileSync' benchmark/auto-resolve/fixtures/F31-cli-seat-rebalance/verifiers/priority-transfer-rollback.js \
+  || ! grep -Fq "assert.strictEqual(result.stderr, '')" benchmark/auto-resolve/fixtures/F31-cli-seat-rebalance/verifiers/priority-transfer-rollback.js \
+  || ! grep -Fq 'assert.deepStrictEqual(parsed, {' benchmark/auto-resolve/fixtures/F31-cli-seat-rebalance/verifiers/priority-transfer-rollback.js \
+  || grep -Fq 'parsed.applied' benchmark/auto-resolve/fixtures/F31-cli-seat-rebalance/verifiers/priority-transfer-rollback.js \
+  || ! grep -Fq 'On success, write exactly one JSON object to stdout and no stderr. Keys: `applied`, `rejected`, `accounts`.' benchmark/auto-resolve/fixtures/F31-cli-seat-rebalance/expected.json \
+  || [ "$(grep -Fc 'On success, write exactly one JSON object to stdout and no stderr. Keys: `applied`, `rejected`, `accounts`.' benchmark/auto-resolve/fixtures/F31-cli-seat-rebalance/expected.json)" -ne 1 ]; then
+  offenders="${offenders}"$'\n'"F31 priority verifier must bind success stderr/no-extra-output contract"
+fi
+if ! grep -Fq '20260512-f32-subscription-renewal-headroom' benchmark/auto-resolve/fixtures/F32-cli-subscription-renewal/NOTES.md \
+  || ! grep -Fq '20260512-f32-subscription-renewal-headroom' benchmark/auto-resolve/BENCHMARK-RESULTS.md \
+  || ! grep -Fq '20260512-f32-subscription-renewal-headroom' benchmark/auto-resolve/README.md \
+  || ! grep -Fq '20260512-f32-subscription-renewal-headroom' benchmark/auto-resolve/run-real-benchmark.md; then
+  offenders="${offenders}"$'\n'"F32 docs must cite the measured headroom rejection before anyone counts it"
+fi
+if ! grep -Fq 'assert.deepEqual(output, {' benchmark/auto-resolve/fixtures/F32-cli-subscription-renewal/verifiers/priority-credit-rollback.js \
+  || ! grep -Fq 'Output row key names and nested `credits` key names match the visible spec exactly, with no aliased or extra keys.' benchmark/auto-resolve/fixtures/F32-cli-subscription-renewal/expected.json \
+  || [ "$(grep -Fc 'Output row key names and nested `credits` key names match the visible spec exactly, with no aliased or extra keys.' benchmark/auto-resolve/fixtures/F32-cli-subscription-renewal/expected.json)" -ne 1 ]; then
+  offenders="${offenders}"$'\n'"F32 priority verifier must own exact output-shape contract without duplicate-error overclaim"
+fi
+if ! grep -Fq 'retired_fixture_exists' benchmark/auto-resolve/scripts/run-headroom-candidate.sh \
+  || ! grep -Fq 'retired_fixture_exists' benchmark/auto-resolve/scripts/run-full-pipeline-pair-candidate.sh \
+  || ! grep -Fq 'fixture is retired and is not rerun by pair-candidate runners' benchmark/auto-resolve/scripts/test-run-headroom-candidate.sh \
+  || ! grep -Fq 'fixture is retired and is not rerun by pair-candidate runners' benchmark/auto-resolve/scripts/test-run-full-pipeline-pair-candidate.sh \
+  || ! grep -Fq 'historical artifact replay' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'historical artifact replay' benchmark/auto-resolve/run-real-benchmark.md; then
+  offenders="${offenders}"$'\n'"pair candidate runners must reject retired fixtures explicitly and docs must reserve retired fixtures for historical artifact replay"
+fi
+if ! grep -Fq 'F16-cli-quote-tax-rules F23-cli-fulfillment-wave F25-cli-cart-promotion-rules' benchmark/auto-resolve/run-real-benchmark.md; then
+  offenders="${offenders}"$'\n'"run-real-benchmark.md examples must use the current measured F16/F23/F25 pair-evidence set"
+fi
+if ! grep -Fq '20260510-f16-f23-f25-combined-proof' benchmark/auto-resolve/BENCHMARK-RESULTS.md \
+  || ! grep -Fq '20260510-f16-f23-f25-combined-proof' benchmark/auto-resolve/README.md \
+  || ! grep -Fq '20260510-f16-f23-f25-combined-proof' benchmark/auto-resolve/run-real-benchmark.md; then
+  offenders="${offenders}"$'\n'"benchmark docs must cite the current F16/F23/F25 pair proof run"
+fi
+if ! grep -Fq 'average pair margin +25.3' benchmark/auto-resolve/BENCHMARK-RESULTS.md \
+  || ! grep -Fq 'average pair margin +25.3' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'average pair margin was `+25.3`' benchmark/auto-resolve/run-real-benchmark.md; then
+  offenders="${offenders}"$'\n'"benchmark docs must cite the F16/F23/F25 average pair margin"
+fi
+if ! grep -Fq 'average solo_claude headroom **8.0**' benchmark/auto-resolve/BENCHMARK-RESULTS.md \
+  || ! grep -Fq 'minimum solo_claude' benchmark/auto-resolve/BENCHMARK-RESULTS.md \
+  || ! grep -Fq '| Fixture | Bare | Solo_claude | Pair (`l2_risk_probes`) | Margin | Pair mode | Wall ratio |' benchmark/auto-resolve/BENCHMARK-RESULTS.md \
+  || ! grep -Fq 'include `bare` headroom and `solo_claude` headroom' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'default 5-point `bare`/`solo_claude` headroom margins' benchmark/auto-resolve/BENCHMARK-RESULTS.md \
+  || ! grep -Fq 'average and minimum `bare`/`solo_claude`' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'average and minimum headroom' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'fixture pass count' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'fixture pass count' benchmark/auto-resolve/run-real-benchmark.md; then
+  offenders="${offenders}"$'\n'"benchmark docs must cite and explain headroom set summaries"
+fi
+if ! grep -Fq 'both baseline arms evidence-complete' \
+  benchmark/auto-resolve/results/20260510-f16-f23-f25-combined-proof/headroom-gate.md \
+  || ! grep -Fq 'both baseline arms evidence-complete' \
+    benchmark/auto-resolve/results/20260510-f16-f23-f25-combined-proof/headroom-gate.json \
+  || ! grep -Fq 'Fixtures passed: 3/3 (minimum required: 3)' \
+    benchmark/auto-resolve/results/20260510-f16-f23-f25-combined-proof/headroom-gate.md \
+  || ! grep -Fq 'Average solo_claude headroom: 8.0' \
+    benchmark/auto-resolve/results/20260510-f16-f23-f25-combined-proof/headroom-gate.md \
+  || ! grep -Fq '"min_solo_headroom": 5' \
+    benchmark/auto-resolve/results/20260510-f16-f23-f25-combined-proof/headroom-gate.json \
+  || ! grep -Fq '| Fixture | Bare | Bare headroom | Solo_claude | Solo_claude headroom | Status | Reason |' \
+    benchmark/auto-resolve/results/20260510-f16-f23-f25-combined-proof/headroom-gate.md \
+  || ! grep -Fq '| F16-cli-quote-tax-rules | 50 | 10 | 75 | 5 | PASS |  |' \
+    benchmark/auto-resolve/results/20260510-f16-f23-f25-combined-proof/headroom-gate.md \
+  || ! grep -Fq '"solo_headroom": 14' \
+    benchmark/auto-resolve/results/20260510-f16-f23-f25-combined-proof/headroom-gate.json; then
+  offenders="${offenders}"$'\n'"tracked F16/F23/F25 headroom report must use current evidence-complete wording and headroom columns"
+fi
+if ! grep -Fq 'l2_risk_probes evidence-clean' \
+  benchmark/auto-resolve/results/20260510-f16-f23-f25-combined-proof/full-pipeline-pair-gate.md \
+  || ! grep -Fq 'l2_risk_probes must be evidence-clean' \
+    benchmark/auto-resolve/results/20260510-f16-f23-f25-combined-proof/full-pipeline-pair-gate.json \
+  || ! grep -Fq 'Fixtures passed: 3/3 (minimum required: 3)' \
+    benchmark/auto-resolve/results/20260510-f16-f23-f25-combined-proof/full-pipeline-pair-gate.md \
+  || ! grep -Fq 'Average pair margin: +25.3' \
+    benchmark/auto-resolve/results/20260510-f16-f23-f25-combined-proof/full-pipeline-pair-gate.md \
+  || ! grep -Fq '"avg_pair_margin": 25.333333333333332' \
+    benchmark/auto-resolve/results/20260510-f16-f23-f25-combined-proof/full-pipeline-pair-gate.json; then
+  offenders="${offenders}"$'\n'"tracked F16/F23/F25 pair report must use current l2_risk_probes evidence-clean wording and average pair margin"
+fi
+if ! grep -Fq 'evidence-complete `bare <= 60`' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'evidence-complete `bare <= 60`' benchmark/auto-resolve/run-real-benchmark.md; then
+  offenders="${offenders}"$'\n'"benchmark docs must describe baseline headroom arms as evidence-complete, not correctness-clean"
+fi
+if ! grep -Fq 'min-bare-headroom' benchmark/auto-resolve/scripts/headroom-gate.py \
+  || ! grep -Fq 'min-solo-headroom' benchmark/auto-resolve/scripts/headroom-gate.py \
+  || ! grep -Fq 'min-bare-headroom' benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py \
+  || ! grep -Fq 'min-solo-headroom' benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py \
+  || ! grep -Fq 'headroom >= 5' benchmark/auto-resolve/results/20260510-f16-f23-f25-combined-proof/headroom-gate.md \
+  || ! grep -Fq 'headroom >= 5' benchmark/auto-resolve/results/20260510-f16-f23-f25-combined-proof/full-pipeline-pair-gate.md \
+  || ! grep -Fq 'default minimum 5-point `bare`/`solo_claude` headroom margin' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'default minimum 5-point `bare`/`solo_claude` headroom margin' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'default 5-point `bare`/`solo_claude` headroom margins' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'default 5-point `bare`/`solo_claude` headroom margins' benchmark/auto-resolve/BENCHMARK-DESIGN.md \
+  || ! grep -Fq 'bare headroom 1 < 5' benchmark/auto-resolve/scripts/test-headroom-gate.sh \
+  || ! grep -Fq 'bare headroom 1 < 5' benchmark/auto-resolve/scripts/test-full-pipeline-pair-gate.sh; then
+  offenders="${offenders}"$'\n'"benchmark headroom gates must reject threshold-fragile candidates with default 5-point baseline headroom margins"
+fi
+if [ -d benchmark/auto-resolve/results/20260511-f28-headroom-smoke-085307/F28-cli-return-authorization ]; then
+  if python3 benchmark/auto-resolve/scripts/headroom-gate.py \
+    --run-id 20260511-f28-headroom-smoke-085307 \
+    --min-fixtures 1 \
+    --out-json /tmp/devlyn-lint-f28-headroom-gate.json \
+    --out-md /tmp/devlyn-lint-f28-headroom-gate.md >/dev/null 2>&1; then
+    offenders="${offenders}"$'\n'"current local F28 headroom artifact must fail the default 5-point headroom margin gate"
+  elif ! grep -Fq 'bare headroom 1 < 5' /tmp/devlyn-lint-f28-headroom-gate.md; then
+    offenders="${offenders}"$'\n'"current local F28 headroom failure must cite bare headroom 1 < 5"
+  fi
+fi
+if [ -d benchmark/auto-resolve/results/20260511-f28-pair-smoke-091021/F28-cli-return-authorization ]; then
+  if python3 benchmark/auto-resolve/scripts/headroom-gate.py \
+    --run-id 20260511-f28-pair-smoke-091021 \
+    --min-fixtures 1 \
+    --out-json /tmp/devlyn-lint-f28-pair-smoke-headroom-gate.json \
+    --out-md /tmp/devlyn-lint-f28-pair-smoke-headroom-gate.md >/dev/null 2>&1; then
+    offenders="${offenders}"$'\n'"current local F28 pair-smoke artifact must fail before pair evidence can be counted"
+  elif ! grep -Fq 'bare score 65 > 60' /tmp/devlyn-lint-f28-pair-smoke-headroom-gate.md; then
+    offenders="${offenders}"$'\n'"current local F28 pair-smoke failure must cite bare score 65 > 60"
+  fi
+fi
+if ! grep -Fq 'startup `Gate:` line' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'startup `Gate:` line' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'headroom gate passed — candidate set accepted' benchmark/auto-resolve/scripts/run-headroom-candidate.sh \
+  || ! grep -Fq 'headroom gate failed — candidate set rejected' benchmark/auto-resolve/scripts/run-headroom-candidate.sh \
+  || ! grep -Fq 'candidate set was accepted or rejected' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'candidate set was accepted or rejected' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'fixture score table with bare score, bare' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'headroom, solo_claude score, solo_claude headroom' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'remaining headroom against' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'pair minus `solo_claude` margin' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'Fixture                         Bare  Solo_claude  Pair  Pair-Solo_claude' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'fixture score table with bare,' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'solo_claude, pair, margin, pair-mode' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'pair above `solo_claude`' benchmark/auto-resolve/BENCHMARK-RESULTS.md \
+  || ! grep -Fq 'Suite average variant-solo_claude margin' benchmark/auto-resolve/BENCHMARK-DESIGN.md \
+  || ! grep -Fq 'legacy `variant`-`bare` (L2-L0)' benchmark/auto-resolve/RUBRIC.md \
+  || ! grep -Fq '`solo_claude`-`bare` measures solo harness value; pair-`solo_claude` measures pair value' benchmark/auto-resolve/README.md \
+  || ! grep -Fq '| fixture | bare | solo_claude | solo_claude-bare |' benchmark/auto-resolve/scripts/test-run-headroom-candidate.sh \
+  || ! grep -Fq '| fixture | bare | solo_claude | solo_claude-bare |' benchmark/auto-resolve/scripts/test-run-full-pipeline-pair-candidate.sh \
+  || ! grep -Fq '| fixture | bare | solo_claude | pair | pair-solo_claude |' benchmark/auto-resolve/scripts/test-run-full-pipeline-pair-candidate.sh \
+  || ! grep -Fq 'startup `Headroom:` and `Pair:` lines' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'startup `Headroom:` / `Pair:` lines' benchmark/auto-resolve/README.md; then
+  offenders="${offenders}"$'\n'"benchmark docs must describe real-run startup gate lines and headroom score columns"
+fi
+if ! grep -Fq 'DEVLYN_BENCHMARK_CLI_SUBCOMMAND: benchmarkMode' bin/devlyn.js \
+  || ! grep -Fq 'npx devlyn-cli benchmark headroom --run-id "$RUN_ID"' benchmark/auto-resolve/scripts/run-headroom-candidate.sh \
+  || ! grep -Fq 'npx devlyn-cli benchmark pair --run-id "$RUN_ID"' benchmark/auto-resolve/scripts/run-full-pipeline-pair-candidate.sh \
+  || ! grep -Fq 'canonical trigger, margin >= +' benchmark/auto-resolve/scripts/run-full-pipeline-pair-candidate.sh \
+  || ! grep -Fq 'canonical trigger, margin >= +5' benchmark/auto-resolve/scripts/test-run-full-pipeline-pair-candidate.sh \
+  || ! grep -Fq 'headroom gate failed — pair arm not executed' benchmark/auto-resolve/scripts/run-full-pipeline-pair-candidate.sh \
+  || ! grep -Fq 'pair gate failed — pair evidence rejected' benchmark/auto-resolve/scripts/run-full-pipeline-pair-candidate.sh \
+  || ! grep -Fq 'headroom gate passed — executing $PAIR_ARM' benchmark/auto-resolve/scripts/run-full-pipeline-pair-candidate.sh \
+  || ! grep -Fq 'pair gate passed — pair evidence accepted' benchmark/auto-resolve/scripts/run-full-pipeline-pair-candidate.sh \
+  || ! grep -Fq 'headroom fails, the runner explicitly says the pair arm was not executed' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'pair gate fails, it explicitly says pair evidence was rejected' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'then that pair evidence was accepted' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'If headroom fails, it reports that the pair arm was not executed' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'pair gate fails, it reports that pair evidence was rejected' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'then that pair evidence was' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'accepted. When launched through' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'the replay `Command:` uses the' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'same package CLI path' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'the replay command uses' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'uses that same' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'package CLI path' benchmark/auto-resolve/run-real-benchmark.md; then
+  offenders="${offenders}"$'\n'"benchmark CLI headroom/pair runs must replay as npx devlyn-cli commands and docs must state that"
+fi
+if ! grep -Fq 'npx devlyn-cli benchmark headroom --min-fixtures 3 F16-cli-quote-tax-rules F23-cli-fulfillment-wave F25-cli-cart-promotion-rules' README.md \
+  || ! grep -Fq 'npx devlyn-cli benchmark frontier --out-md /tmp/devlyn-pair-frontier.md' README.md \
+  || ! grep -Fq 'npx devlyn-cli benchmark pair --min-fixtures 3 --max-pair-solo-wall-ratio 3 F16-cli-quote-tax-rules F23-cli-fulfillment-wave F25-cli-cart-promotion-rules' README.md \
+  || ! grep -Fq 'average pair margin' README.md \
+  || ! grep -Fq 'default 5-point `bare`/`solo_claude` headroom margins' README.md \
+  || ! grep -Fq 'Add `--dry-run` to either score runner' README.md \
+  || ! grep -Fq 'fixture count, and the replay command' README.md \
+  || ! grep -Fq 'Dry-runs' README.md \
+  || ! grep -Fq 'and lint prove wiring only; real score claims must cite the run id and fixture' README.md; then
+  offenders="${offenders}"$'\n'"README.md must expose score-focused benchmark headroom/pair CLI paths and distinguish wiring checks from real scores"
+fi
+if ! grep -Fq 'npx devlyn-cli benchmark pair --min-fixtures 3 --max-pair-solo-wall-ratio 3 F16-cli-quote-tax-rules F23-cli-fulfillment-wave F25-cli-cart-promotion-rules' bin/devlyn.js \
+  || ! grep -Fq -- '--max-pair-solo-wall-ratio N  default: 3' bin/devlyn.js \
+  || ! grep -Fq 'MAX_PAIR_SOLO_WALL_RATIO=3' benchmark/auto-resolve/scripts/run-full-pipeline-pair-candidate.sh \
+  || ! grep -Fq 'parser.add_argument("--max-pair-solo-wall-ratio", type=positive_float, default=3.0)' benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py \
+  || ! grep -Fq '"max_observed_pair_solo_wall_ratio": max(ratios) if ratios else None' benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py \
+  || ! grep -Fq 'Allowed pair/solo wall ratio' benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py \
+  || ! grep -Fq 'Maximum observed pair/solo wall ratio' benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py \
+  || ! grep -Fq 'Maximum observed pair/solo wall ratio: 2.00x' benchmark/auto-resolve/scripts/test-full-pipeline-pair-gate.sh \
+  || ! grep -Fq 'separates the allowed pair/solo wall ratio from the maximum observed pair/solo wall ratio' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'python3 "$GATE" --results-root "$TMP_DIR" --run-id slow-pair' benchmark/auto-resolve/scripts/test-full-pipeline-pair-gate.sh \
+  || ! grep -Fq -- '--min-fixtures 3' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq -- '--min-fixtures 3' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'average pair margin' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'average pair margin' benchmark/auto-resolve/README.md \
+  || ! grep -Fq -- '--dry-run          validate args/fixtures and print replay command only' bin/devlyn.js \
+  || ! grep -Fq 'does not produce' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'scores. When showing' benchmark/auto-resolve/run-real-benchmark.md; then
+  offenders="${offenders}"$'\n'"benchmark pair examples must explicitly gate the current F16/F23/F25 proof set with --min-fixtures 3"
+fi
+if ! grep -Fq 'use 3 for F16/F23/F25 proof reruns; audit requires 4 passing evidence rows' bin/devlyn.js \
+  || ! grep -Fq 'use 3 for F16/F23/F25 proof reruns; audit requires 4 passing evidence rows' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh; then
+  offenders="${offenders}"$'\n'"benchmark CLI help must distinguish proof reruns from the four-row release audit"
+fi
+if ! grep -Fq 'judge blind mapping missing' benchmark/auto-resolve/scripts/headroom-gate.py \
+  || ! grep -Fq 'judge blind mapping missing' benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py \
+  || ! grep -Fq 'blind_mapping_arm_missing' benchmark/auto-resolve/scripts/compile-report.py \
+  || ! grep -Fq 'def trusted_winner()' benchmark/auto-resolve/scripts/compile-report.py \
+  || ! grep -Fq 'raw_findings_by_arm = judge.get("findings_by_arm")' benchmark/auto-resolve/scripts/compile-report.py \
+  || ! grep -Fq 'def critical_findings_for(arm: str)' benchmark/auto-resolve/scripts/compile-report.py \
+  || ! grep -Fq 'def exact_bool(value)' benchmark/auto-resolve/scripts/compile-report.py \
+  || ! grep -Fq '"malformed_boolean_fields": malformed_boolean_fields' benchmark/auto-resolve/scripts/compile-report.py \
+  || ! grep -Fq '"dq_judge_malformed": judge_dq_malformed' benchmark/auto-resolve/scripts/compile-report.py \
+  || ! grep -Fq 'def bool_or_none(value)' benchmark/auto-resolve/scripts/ship-gate.py \
+  || ! grep -Fq 'l1_dq_by_fixture: dict[str, bool]' benchmark/auto-resolve/scripts/ship-gate.py \
+  || ! grep -Fq 'summary arms_present malformed' benchmark/auto-resolve/scripts/ship-gate.py \
+  || ! grep -Fq 'summary arms_present.solo_claude malformed' benchmark/auto-resolve/scripts/ship-gate.py \
+  || ! grep -Fq 'summary margins_avg malformed' benchmark/auto-resolve/scripts/ship-gate.py \
+  || ! grep -Fq 'raw_findings_letters = chosen.get("critical_findings")' benchmark/auto-resolve/scripts/judge.sh \
+  || ! grep -Fq 'raw_findings_letters = chosen.get("critical_findings")' benchmark/auto-resolve/scripts/judge-opus-pass.sh \
+  || ! grep -Fq 'raw_validation = judge.get("_axis_validation")' benchmark/auto-resolve/scripts/compile-report.py \
+  || ! grep -Fq 'raw_validation = judge.get("_axis_validation")' benchmark/auto-resolve/scripts/headroom-gate.py \
+  || ! grep -Fq 'raw_validation = judge.get("_axis_validation")' benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py \
+  || ! grep -Fq 'raw_scores_by_arm = judge.get("scores_by_arm")' benchmark/auto-resolve/scripts/compile-report.py \
+  || ! grep -Fq 'from pair_evidence_contract import is_score, is_strict_number' benchmark/auto-resolve/scripts/compile-report.py \
+  || ! grep -Fq 'loads_strict_json_object(path.read_text())' benchmark/auto-resolve/scripts/compile-report.py \
+  || ! grep -Fq 'parse_constant=reject_json_constant' benchmark/auto-resolve/scripts/ship-gate.py \
+  || ! grep -Fq 'if arm in mapped_arms and is_score(score)' benchmark/auto-resolve/scripts/compile-report.py \
+  || ! grep -Fq 'def strict_number(value)' benchmark/auto-resolve/scripts/compile-report.py \
+  || ! grep -Fq 'raw_scores = judge.get("scores_by_arm")' benchmark/auto-resolve/scripts/headroom-gate.py \
+  || ! grep -Fq 'raw_scores = judge.get("scores_by_arm")' benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py \
+  || ! grep -Fq 'from pair_evidence_contract import is_score' benchmark/auto-resolve/scripts/headroom-gate.py \
+  || ! grep -Fq 'return value if is_score(value) else None' benchmark/auto-resolve/scripts/headroom-gate.py \
+  || ! grep -Fq 'raw_scores = judge.get("scores_by_arm")' benchmark/auto-resolve/scripts/build-pair-eligible-manifest.py \
+  || ! grep -Fq 'from pair_evidence_contract import is_score' benchmark/auto-resolve/scripts/build-pair-eligible-manifest.py \
+  || ! grep -Fq 'def exact_bool(value: object)' benchmark/auto-resolve/scripts/build-pair-eligible-manifest.py \
+  || ! grep -Fq 'def disqualifier_flag(value: object' benchmark/auto-resolve/scripts/build-pair-eligible-manifest.py \
+  || ! grep -Fq 'if not is_score(solo) or not is_score(bare):' benchmark/auto-resolve/scripts/build-pair-eligible-manifest.py \
+  || ! grep -Fq 'if is_score(score):' benchmark/auto-resolve/scripts/build-pair-eligible-manifest.py \
+  || ! grep -Fq 'return legacy if is_score(legacy) else None' benchmark/auto-resolve/scripts/build-pair-eligible-manifest.py \
+  || ! grep -Fq 'raw_scores = judge.get("scores_by_arm")' benchmark/auto-resolve/scripts/iter-0033c-l1-summary.py \
+  || ! grep -Fq 'raw_scores = judge.get("scores_by_arm")' benchmark/auto-resolve/scripts/iter-0033c-compare.py \
+  || ! grep -Fq 'from pair_evidence_contract import (' benchmark/auto-resolve/scripts/iter-0033c-compare.py \
+  || ! grep -Fq 'is_score,' benchmark/auto-resolve/scripts/iter-0033c-compare.py \
+  || ! grep -Fq 'is_strict_number,' benchmark/auto-resolve/scripts/iter-0033c-compare.py \
+  || ! grep -Fq 'loads_strict_json_object(p.read_text())' benchmark/auto-resolve/scripts/iter-0033c-compare.py \
+  || ! grep -Fq 'parse_constant=reject_json_constant' benchmark/auto-resolve/scripts/iter-0033c-compare.py \
+  || ! grep -Fq 'PAIR_VERDICTS = {"PASS", "PASS_WITH_ISSUES", "NEEDS_WORK", "BLOCKED", "FAIL"}' benchmark/auto-resolve/scripts/iter-0033c-compare.py \
+  || ! grep -Fq 'def is_pair_judge_verdict(value: object) -> bool:' benchmark/auto-resolve/scripts/iter-0033c-compare.py \
+  || ! grep -Fq 'def exact_bool(value: object)' benchmark/auto-resolve/scripts/iter-0033c-compare.py \
+  || ! grep -Fq 'def bool_flag(value: object' benchmark/auto-resolve/scripts/iter-0033c-compare.py \
+  || ! grep -Fq 'if is_score(sba.get(arm)):' benchmark/auto-resolve/scripts/iter-0033c-compare.py \
+  || ! grep -Fq 'return legacy if is_score(legacy) else None' benchmark/auto-resolve/scripts/iter-0033c-compare.py \
+  || ! grep -Fq 'def strict_elapsed_seconds' benchmark/auto-resolve/scripts/iter-0033c-compare.py \
+  || ! grep -Fq 'def timeout_flag(result: dict | None) -> bool:' benchmark/auto-resolve/scripts/iter-0033c-compare.py \
+  || ! grep -Fq 'raw_scores = judge.get("scores_by_arm")' benchmark/auto-resolve/scripts/judge-opus-pass.sh \
+  || ! grep -Fq 'raw_dq_by_arm = judge.get("disqualifiers_by_arm")' benchmark/auto-resolve/scripts/headroom-gate.py \
+  || ! grep -Fq 'raw_dq_by_arm = judge.get("disqualifiers_by_arm")' benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py \
+  || ! grep -Fq 'raw_dq_by_arm = judge.get("disqualifiers_by_arm")' benchmark/auto-resolve/scripts/compile-report.py \
+  || ! grep -Fq 'raw_by_arm = judge.get("disqualifiers_by_arm")' benchmark/auto-resolve/scripts/build-pair-eligible-manifest.py \
+  || ! grep -Fq 'raw_dba = judge.get("disqualifiers_by_arm")' benchmark/auto-resolve/scripts/iter-0033c-compare.py \
+  || ! grep -Fq 'raw_legacy = judge.get("disqualifiers")' benchmark/auto-resolve/scripts/build-pair-eligible-manifest.py \
+  || ! grep -Fq 'raw_dqs = judge.get("disqualifiers")' benchmark/auto-resolve/scripts/iter-0033c-compare.py \
+  || ! grep -Fq 'def is_score(value):' benchmark/auto-resolve/scripts/judge.sh \
+  || ! grep -Fq 'invalid judge score value(s)' benchmark/auto-resolve/scripts/judge.sh \
+  || ! grep -Fq 'invalid judge disqualifier value(s)' benchmark/auto-resolve/scripts/judge.sh \
+  || ! grep -Fq 'invalid opus disqualifier value(s)' benchmark/auto-resolve/scripts/judge-opus-pass.sh \
+  || ! grep -Fq 'if arm is not None and key in chosen and is_score(chosen[key]):' benchmark/auto-resolve/scripts/judge.sh \
+  || ! grep -Fq 'raw_dq_letters = chosen.get("disqualifiers")' benchmark/auto-resolve/scripts/judge.sh \
+  || ! grep -Fq 'raw_dq_letters = chosen.get("disqualifiers")' benchmark/auto-resolve/scripts/judge-opus-pass.sh \
+  || ! grep -Fq 'scores_by_arm` alone is not evidence' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'without the blind slot mapping is not score evidence' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'absent from the blind mapping is not score evidence' benchmark/auto-resolve/scripts/iter-0033c-compare.py \
+  || ! grep -Fq 'wrong-pair-mapping' benchmark/auto-resolve/scripts/test-full-pipeline-pair-gate.sh \
+  || ! grep -Fq 'from pair_evidence_contract import (' benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py \
+  || ! grep -Fq 'ALLOWED_PAIR_ARMS,' benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py \
+  || ! grep -Fq 'loads_strict_json_object,' benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py \
+  || ! grep -Fq 'def bool_flag_failure(value: Any, true_reason: str, malformed_reason: str) -> str | None:' benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py \
+  || ! grep -Fq 'def bool_flag_failure(value: object, true_reason: str, malformed_reason: str) -> str | None:' benchmark/auto-resolve/scripts/headroom-gate.py \
+  || ! grep -Fq 'return value if is_score(value) else None' benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py \
+  || ! grep -Fq 'if not is_strict_number(pair_elapsed) or not is_strict_number(solo_elapsed):' benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py \
+  || ! grep -Fq 'return is_strict_number(value) and value >= 1.0' benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py \
+  || ! grep -Fq 'value must be finite and > 0' benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py \
+  || ! grep -Fq 'overrange-score' benchmark/auto-resolve/scripts/test-full-pipeline-pair-gate.sh \
+  || ! grep -Fq 'boolean-score' benchmark/auto-resolve/scripts/test-full-pipeline-pair-gate.sh \
+  || ! grep -Fq 'boolean-wall-time' benchmark/auto-resolve/scripts/test-full-pipeline-pair-gate.sh \
+  || ! grep -Fq 'invalid-max-wall-ratio' benchmark/auto-resolve/scripts/test-full-pipeline-pair-gate.sh \
+  || ! grep -Fq 'boolean-pair-verify-score' benchmark/auto-resolve/scripts/test-full-pipeline-pair-gate.sh \
+  || ! grep -Fq 'malformed-pair-bool' benchmark/auto-resolve/scripts/test-full-pipeline-pair-gate.sh \
+  || ! grep -Fq 'malformed-judge-bool' benchmark/auto-resolve/scripts/test-full-pipeline-pair-gate.sh \
+  || ! grep -Fq 'wrong-mapping' benchmark/auto-resolve/scripts/test-headroom-gate.sh \
+  || ! grep -Fq 'overrange-score' benchmark/auto-resolve/scripts/test-headroom-gate.sh \
+  || ! grep -Fq 'boolean-score' benchmark/auto-resolve/scripts/test-headroom-gate.sh \
+  || ! grep -Fq 'malformed-bare-bool' benchmark/auto-resolve/scripts/test-headroom-gate.sh \
+  || ! grep -Fq 'bare judge disqualifier malformed' benchmark/auto-resolve/scripts/test-headroom-gate.sh \
+  || ! grep -Fq 'variant-mapping-disqualifies' benchmark/auto-resolve/scripts/test-ship-gate.sh \
+  || ! grep -Fq 'solo-mapping-disqualifies' benchmark/auto-resolve/scripts/test-ship-gate.sh \
+  || ! grep -Fq 'stale-margin' benchmark/auto-resolve/scripts/test-ship-gate.sh \
+  || ! grep -Fq 'stale judge margins must be recomputed from trusted scores' benchmark/auto-resolve/scripts/test-ship-gate.sh \
+  || ! grep -Fq 'winner without blind-mapped trusted score must be null' benchmark/auto-resolve/scripts/test-ship-gate.sh \
+  || ! grep -Fq 'malformed scores_by_arm must not expose' benchmark/auto-resolve/scripts/test-ship-gate.sh \
+  || ! grep -Fq 'out-of-range scores_by_arm must not expose variant score' benchmark/auto-resolve/scripts/test-ship-gate.sh \
+  || ! grep -Fq 'boolean scores_by_arm must not expose solo score' benchmark/auto-resolve/scripts/test-ship-gate.sh \
+  || ! grep -Fq 'boolean result numeric fields must not appear in compile summary' benchmark/auto-resolve/scripts/test-ship-gate.sh \
+  || ! grep -Fq 'malformed-result-bool' benchmark/auto-resolve/scripts/test-ship-gate.sh \
+  || ! grep -Fq 'malformed-judge-bool' benchmark/auto-resolve/scripts/test-ship-gate.sh \
+  || ! grep -Fq 'malformed-l1-dq-summary' benchmark/auto-resolve/scripts/test-ship-gate.sh \
+  || ! grep -Fq 'malformed-arms-present-wrapper' benchmark/auto-resolve/scripts/test-ship-gate.sh \
+  || ! grep -Fq 'malformed-arms-present' benchmark/auto-resolve/scripts/test-ship-gate.sh \
+  || ! grep -Fq 'malformed-margins-avg-wrapper' benchmark/auto-resolve/scripts/test-ship-gate.sh \
+  || ! grep -Fq 'F1 L1 disqualifier malformed' benchmark/auto-resolve/scripts/test-ship-gate.sh \
+  || ! grep -Fq 'overrange C1 row fields must not promote F17' benchmark/auto-resolve/scripts/test-build-pair-eligible-manifest.sh \
+  || ! grep -Fq 'string C1 disqualifier must not promote F18' benchmark/auto-resolve/scripts/test-build-pair-eligible-manifest.sh \
+  || ! grep -Fq 'f9-overrange-scores' benchmark/auto-resolve/scripts/test-build-pair-eligible-manifest.sh \
+  || ! grep -Fq 'f9-boolean-scores' benchmark/auto-resolve/scripts/test-build-pair-eligible-manifest.sh \
+  || ! grep -Fq 'f9-string-dq-entry' benchmark/auto-resolve/scripts/test-build-pair-eligible-manifest.sh \
+  || ! grep -Fq 'overrange-scores-results' benchmark/auto-resolve/scripts/test-iter-0033c-compare.sh \
+  || ! grep -Fq 'boolean-scores-results' benchmark/auto-resolve/scripts/test-iter-0033c-compare.sh \
+  || ! grep -Fq 'boolean-wall-results' benchmark/auto-resolve/scripts/test-iter-0033c-compare.sh \
+  || ! grep -Fq 'nan-wall-results' benchmark/auto-resolve/scripts/test-iter-0033c-compare.sh \
+  || ! grep -Fq 'string-timeout-results' benchmark/auto-resolve/scripts/test-iter-0033c-compare.sh \
+  || ! grep -Fq 'malformed-pair-state-results' benchmark/auto-resolve/scripts/test-iter-0033c-compare.sh \
+  || ! grep -Fq 'non-list finding entry must become a one-item list' benchmark/auto-resolve/scripts/test-ship-gate.sh \
+  || ! grep -Fq 'non-dict findings_by_arm must be ignored' benchmark/auto-resolve/scripts/test-ship-gate.sh \
+  || ! grep -Fq 'non-dict _axis_validation wrapper must not crash' benchmark/auto-resolve/scripts/test-ship-gate.sh \
+  || ! grep -Fq 'load_dict_json' benchmark/auto-resolve/scripts/compile-report.py \
+  || ! grep -Fq 'measurement invalid: malformed summary.json' benchmark/auto-resolve/scripts/test-ship-gate.sh \
+  || ! grep -Fq 'nan-summary' benchmark/auto-resolve/scripts/test-ship-gate.sh \
+  || ! grep -Fq 'summary rows contain non-object entries' benchmark/auto-resolve/scripts/test-ship-gate.sh \
+  || ! grep -Fq 'malformed-summary-field-types' benchmark/auto-resolve/scripts/test-ship-gate.sh \
+  || ! grep -Fq 'variant axis count malformed' benchmark/auto-resolve/scripts/test-ship-gate.sh \
+  || ! grep -Fq 'result.json malformed' benchmark/auto-resolve/scripts/test-headroom-gate.sh \
+  || ! grep -Fq 'result.json malformed' benchmark/auto-resolve/scripts/test-full-pipeline-pair-gate.sh \
+  || ! grep -Fq 'non-dict variant result.json must fail closed' benchmark/auto-resolve/scripts/test-ship-gate.sh \
+  || ! grep -Fq 'NaN variant result.json must fail closed as a disqualifier' benchmark/auto-resolve/scripts/test-ship-gate.sh \
+  || ! grep -Fq '("solo_claude (L1)", "solo_claude")' benchmark/auto-resolve/scripts/compile-report.py \
+  || ! grep -Fq '| Fixture | Category | variant (L2) | solo_claude (L1) | bare (L0) | variant-bare | solo_claude-bare | variant-solo_claude | Winner | Wall variant/solo_claude/bare | Wall variant/solo_claude | Wall variant/bare |' benchmark/auto-resolve/scripts/compile-report.py \
+  || ! grep -Fq '| Fixture | Category | variant (L2) | solo_claude (L1) | bare (L0) | variant-bare | solo_claude-bare | variant-solo_claude | Winner | Wall variant/solo_claude/bare | Wall variant/solo_claude | Wall variant/bare |' benchmark/auto-resolve/scripts/test-ship-gate.sh \
+  || ! grep -Fq '**solo_claude (L1):**' benchmark/auto-resolve/scripts/test-ship-gate.sh \
+  || ! grep -Fq 'variant (L2) vs solo_claude (L1) margin avg' benchmark/auto-resolve/scripts/compile-report.py \
+  || ! grep -Fq '**variant (L2) vs solo_claude (L1) margin avg:** +10.0' benchmark/auto-resolve/scripts/test-ship-gate.sh \
+  || ! grep -Fq 'Wall ratio variant (L2) / solo_claude (L1)' benchmark/auto-resolve/scripts/compile-report.py \
+  || ! grep -Fq '**Wall ratio variant (L2) / solo_claude (L1):** 1.0x' benchmark/auto-resolve/scripts/test-ship-gate.sh \
+  || ! grep -Fq 'malformed-scores' benchmark/auto-resolve/scripts/test-headroom-gate.sh \
+  || ! grep -Fq 'malformed-scores' benchmark/auto-resolve/scripts/test-full-pipeline-pair-gate.sh \
+  || ! grep -Fq 'malformed-dq-entry' benchmark/auto-resolve/scripts/test-headroom-gate.sh \
+  || ! grep -Fq 'malformed-dq-entry' benchmark/auto-resolve/scripts/test-full-pipeline-pair-gate.sh \
+  || ! grep -Fq 'malformed-axis-wrapper' benchmark/auto-resolve/scripts/test-headroom-gate.sh \
+  || ! grep -Fq 'malformed-axis-wrapper' benchmark/auto-resolve/scripts/test-full-pipeline-pair-gate.sh \
+  || ! grep -Fq 'fixture-nan-metadata' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'fixture-nan-expected' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'c1-summary malformed: expected object' benchmark/auto-resolve/scripts/test-build-pair-eligible-manifest.sh \
+  || ! grep -Fq 'f9-judge malformed: expected object' benchmark/auto-resolve/scripts/test-build-pair-eligible-manifest.sh \
+  || ! grep -Fq 'malformed C1 row fields must not promote F1' benchmark/auto-resolve/scripts/test-build-pair-eligible-manifest.sh \
+  || ! grep -Fq 'F9 must not be included when scores_by_arm is malformed' benchmark/auto-resolve/scripts/test-build-pair-eligible-manifest.sh \
+  || ! grep -Fq 'truthy malformed disqualifier entry must exclude F9' benchmark/auto-resolve/scripts/test-build-pair-eligible-manifest.sh \
+  || ! grep -Fq 'disqualifiers": ["not", "a", "dict"]' benchmark/auto-resolve/scripts/test-build-pair-eligible-manifest.sh \
+  || ! grep -Fq 'bad-mapping' benchmark/auto-resolve/scripts/test-iter-0033c-compare.sh \
+  || ! grep -Fq 'write_fixture_with_malformed_dq_entry' benchmark/auto-resolve/scripts/test-iter-0033c-compare.sh \
+  || ! grep -Fq 'write_fixture_with_malformed_legacy_dq' benchmark/auto-resolve/scripts/test-iter-0033c-compare.sh \
+  || ! grep -Fq 'write_fixture_with_string_dq_entry' benchmark/auto-resolve/scripts/test-iter-0033c-compare.sh \
+  || ! grep -Fq 'write_fixture_with_malformed_result' benchmark/auto-resolve/scripts/test-iter-0033c-compare.sh \
+  || ! grep -Fq 'disqualifiers": ["not", "a", "dict"]' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'malformed compare.json for malformed-compare' benchmark/auto-resolve/scripts/test-frozen-verify-gate.sh \
+  || ! grep -Fq 'malformed compare.json for nan-compare: invalid JSON' benchmark/auto-resolve/scripts/test-frozen-verify-gate.sh \
+  || ! grep -Fq 'malformed-compare-sections' benchmark/auto-resolve/scripts/test-frozen-verify-gate.sh \
+  || ! grep -Fq 'malformed-verdict-fields' benchmark/auto-resolve/scripts/test-frozen-verify-gate.sh \
+  || ! grep -Fq 'malformed-elapsed-fields' benchmark/auto-resolve/scripts/test-frozen-verify-gate.sh \
+  || ! grep -Fq 'string-pair-mode' benchmark/auto-resolve/scripts/test-frozen-verify-gate.sh \
+  || ! grep -Fq 'def strict_positive_number(value):' benchmark/auto-resolve/scripts/run-frozen-verify-pair.sh \
+  || ! grep -Fq 'math.isfinite(value)' benchmark/auto-resolve/scripts/run-frozen-verify-pair.sh \
+  || ! grep -Fq 'def strict_nonnegative_int(value):' benchmark/auto-resolve/scripts/run-frozen-verify-pair.sh \
+  || ! grep -Fq 'def summary_findings_count(data):' benchmark/auto-resolve/scripts/run-frozen-verify-pair.sh \
+  || ! grep -Fq 'def severity_count_sum(data):' benchmark/auto-resolve/scripts/run-frozen-verify-pair.sh \
+  || ! grep -Fq 'def strict_greater(left, right):' benchmark/auto-resolve/scripts/run-frozen-verify-pair.sh \
+  || ! grep -Fq 'metadata = loads_strict_json_object(pathlib.Path(sys.argv[1]).read_text())' benchmark/auto-resolve/scripts/run-frozen-verify-pair.sh \
+  || ! grep -Fq 'metadata timeout_seconds must be a positive integer' benchmark/auto-resolve/scripts/run-frozen-verify-pair.sh \
+  || ! grep -Fq 'state = as_dict(loads_strict_json_object(state_path.read_text())) if state_path.is_file() else {}' benchmark/auto-resolve/scripts/run-frozen-verify-pair.sh \
+  || ! grep -Fq 'json.loads(line, parse_constant=reject_json_constant)' benchmark/auto-resolve/scripts/run-frozen-verify-pair.sh \
+  || ! grep -Fq 'summary = loads_strict_json_object(pathlib.Path(sys.argv[1]).read_text())' benchmark/auto-resolve/scripts/run-frozen-verify-pair.sh \
+  || ! grep -Fq 'expected = loads_strict_json_object(pathlib.Path(sys.argv[1]).read_text())' benchmark/auto-resolve/scripts/run-frozen-verify-pair.sh \
+  || ! grep -Fq 'out[arm] = loads_strict_json_object(path.read_text()) if path.is_file() else {"missing": True}' benchmark/auto-resolve/scripts/run-frozen-verify-pair.sh \
+  || ! grep -Fq 'def as_dict(value):' benchmark/auto-resolve/scripts/run-fixture.sh \
+  || ! grep -Fq 'metadata = loads_strict_json_object(pathlib.Path(sys.argv[1]).read_text())' benchmark/auto-resolve/scripts/run-fixture.sh \
+  || ! grep -Fq 'metadata timeout_seconds must be a positive integer' benchmark/auto-resolve/scripts/run-fixture.sh \
+  || ! grep -Fq 'expected = loads_strict_json_object(pathlib.Path(sys.argv[1]).read_text())' benchmark/auto-resolve/scripts/run-fixture.sh \
+  || ! grep -Fq 'verify = as_dict(loads_strict_json_object(pathlib.Path(result_dir, "verify.json").read_text()))' benchmark/auto-resolve/scripts/run-fixture.sh \
+  || ! grep -Fq 'loads_strict_json_object(pathlib.Path(result_dir, "timing.json").read_text())' benchmark/auto-resolve/scripts/run-fixture.sh \
+  || ! grep -Fq 'loads_strict_json_object(pathlib.Path(state_path).read_text())' benchmark/auto-resolve/scripts/run-fixture.sh \
+  || ! grep -Fq 'phases = as_dict(state.get("phases"))' benchmark/auto-resolve/scripts/run-fixture.sh \
+  || ! grep -Fq 'legacy_verify = as_dict(state.get("verify"))' benchmark/auto-resolve/scripts/run-fixture.sh \
+  || ! grep -Fq 'data = raw_oracle' benchmark/auto-resolve/scripts/run-fixture.sh \
+  || ! grep -Fq 'oracle artifact malformed or unreadable' benchmark/auto-resolve/scripts/run-fixture.sh \
+  || ! grep -Fq '"type": "oracle-error"' benchmark/auto-resolve/scripts/run-fixture.sh \
+  || ! grep -Fq 'verify["oracle_disqualifier"] = True' benchmark/auto-resolve/scripts/run-fixture.sh \
+  || ! grep -Fq 'findings = raw_findings if isinstance(raw_findings, list) else []' benchmark/auto-resolve/scripts/run-fixture.sh \
+  || ! grep -Fq 'if not isinstance(finding, dict):' benchmark/auto-resolve/scripts/run-fixture.sh \
+  || ! grep -Fq 'def as_dict(value):' benchmark/auto-resolve/scripts/run-frozen-verify-pair.sh \
+  || ! grep -Fq 'phases = as_dict(state.get("phases"))' benchmark/auto-resolve/scripts/run-frozen-verify-pair.sh \
+  || ! grep -Fq 'legacy_verify = as_dict(state.get("verify"))' benchmark/auto-resolve/scripts/run-frozen-verify-pair.sh \
+  || ! grep -Fq 'PAIR_VERDICTS = {"PASS", "PASS_WITH_ISSUES", "NEEDS_WORK", "BLOCKED", "FAIL"}' benchmark/auto-resolve/scripts/run-fixture.sh \
+  || ! grep -Fq 'def has_pair_judge_verdict(sub_verdicts):' benchmark/auto-resolve/scripts/run-fixture.sh \
+  || ! grep -Fq 'PAIR_VERDICTS = {"PASS", "PASS_WITH_ISSUES", "NEEDS_WORK", "BLOCKED", "FAIL"}' benchmark/auto-resolve/scripts/run-frozen-verify-pair.sh \
+  || ! grep -Fq 'def has_pair_judge_verdict(sub_verdicts):' benchmark/auto-resolve/scripts/run-frozen-verify-pair.sh \
+  || ! grep -Fq 'or verify_phase.get("pair_mode") is True' benchmark/auto-resolve/scripts/run-fixture.sh \
+  || ! grep -Fq 'has_pair_judge_verdict(sub_verdicts) or verify_phase.get("pair_mode") is True' benchmark/auto-resolve/scripts/run-fixture.sh \
+  || ! grep -Fq 'has_pair_judge_verdict(sub_verdicts)' benchmark/auto-resolve/scripts/run-frozen-verify-pair.sh \
+  || ! grep -Fq 'raw_pair_sub = pair.get("sub_verdicts")' benchmark/auto-resolve/scripts/run-frozen-verify-pair.sh \
+  || ! grep -Fq 'raw_pair_trigger = pair.get("pair_trigger")' benchmark/auto-resolve/scripts/run-frozen-verify-pair.sh \
+  || ! grep -Fq 'pair_trigger = raw_pair_trigger if isinstance(raw_pair_trigger, dict) else {}' benchmark/auto-resolve/scripts/run-frozen-verify-pair.sh \
+  || ! grep -Fq 'pair_mode_true = pair.get("pair_mode") is True' benchmark/auto-resolve/scripts/run-frozen-verify-pair.sh \
+  || ! grep -Fq 'def fmt_trigger_reasons(value):' benchmark/auto-resolve/scripts/run-frozen-verify-pair.sh \
+  || ! grep -Fq "| Arm | Verdict | Pair mode | Triggers | Findings | Elapsed | Invoke exit | Failure |" benchmark/auto-resolve/scripts/run-frozen-verify-pair.sh \
+  || ! grep -Fq '| Arm | Verdict | Pair mode | Triggers | Findings | Elapsed | Invoke exit | Failure |' benchmark/auto-resolve/scripts/test-swebench-frozen-case.sh \
+  || ! grep -Fq 'pair_mode = pair.get("pair_mode") is True' benchmark/auto-resolve/scripts/frozen-verify-gate.py \
+  || ! grep -Fq 'parse_constant=reject_json_constant' benchmark/auto-resolve/scripts/frozen-verify-gate.py \
+  || ! grep -Fq 'loads_strict_json_object(path.read_text())' benchmark/auto-resolve/scripts/check-f9-artifacts.py \
+  || ! grep -Fq 'invalid JSON numeric constant: NaN' benchmark/auto-resolve/scripts/test-check-f9-artifacts.sh \
+  || ! grep -Fq 'loads_strict_json_object(path.read_text(encoding="utf8"))' benchmark/auto-resolve/scripts/swebench-frozen-matrix.py \
+  || ! grep -Fq 'def pair_trigger_failures(' benchmark/auto-resolve/scripts/swebench-frozen-matrix.py \
+  || ! grep -Fq 'def pair_trigger_reasons(pair: dict[str, Any]) -> list[str]:' benchmark/auto-resolve/scripts/swebench-frozen-matrix.py \
+  || ! grep -Fq 'pair_trigger_eligible(pair)' benchmark/auto-resolve/scripts/swebench-frozen-matrix.py \
+  || ! grep -Fq '"pair_trigger_has_canonical_reason": has_canonical_pair_trigger_reason(trigger_reasons)' benchmark/auto-resolve/scripts/swebench-frozen-matrix.py \
+  || ! grep -Fq '"pair_trigger_has_canonical_reason": true' benchmark/auto-resolve/scripts/test-swebench-frozen-case.sh \
+  || ! grep -Fq 'failed attempt: pair trigger contract: ' benchmark/auto-resolve/scripts/swebench-frozen-matrix.py \
+  || ! grep -Fq 'has_known_pair_trigger_reason(reasons)' benchmark/auto-resolve/scripts/swebench-frozen-matrix.py \
+  || ! grep -Fq 'all_known_pair_trigger_reasons(trigger["reasons"])' benchmark/auto-resolve/scripts/swebench-frozen-matrix.py \
+  || ! grep -Fq 'has_canonical_pair_trigger_reason(trigger["reasons"])' benchmark/auto-resolve/scripts/swebench-frozen-matrix.py \
+  || ! grep -Fq 'path_has_actionable_solo_headroom_hypothesis(fixture_spec)' benchmark/auto-resolve/scripts/swebench-frozen-matrix.py \
+  || ! grep -Fq 'pair_trigger reasons missing known trigger reason' benchmark/auto-resolve/scripts/swebench-frozen-matrix.py \
+  || ! grep -Fq 'pair_trigger reasons contain unknown trigger reason' benchmark/auto-resolve/scripts/swebench-frozen-matrix.py \
+  || ! grep -Fq 'pair_trigger reasons missing canonical trigger reason' benchmark/auto-resolve/scripts/swebench-frozen-matrix.py \
+  || ! grep -Fq 'pair_trigger missing spec.solo_headroom_hypothesis' benchmark/auto-resolve/scripts/swebench-frozen-matrix.py \
+  || ! grep -Fq -- '--require-hypothesis-trigger requires --fixtures-root' benchmark/auto-resolve/scripts/swebench-frozen-matrix.py \
+  || ! grep -Fq 'swebench-missing-hypothesis-trigger-test' benchmark/auto-resolve/scripts/test-swebench-frozen-case.sh \
+  || ! grep -Fq -- '--require-hypothesis-trigger requires --fixtures-root' benchmark/auto-resolve/scripts/test-swebench-frozen-case.sh \
+  || ! grep -Fq 'has_known_pair_trigger_reason(reasons)' benchmark/auto-resolve/scripts/frozen-verify-gate.py \
+  || ! grep -Fq 'all_known_pair_trigger_reasons(reasons)' benchmark/auto-resolve/scripts/frozen-verify-gate.py \
+  || ! grep -Fq 'has_canonical_pair_trigger_reason(reasons)' benchmark/auto-resolve/scripts/frozen-verify-gate.py \
+  || ! grep -Fq 'def pair_trigger_reasons(pair: dict[str, Any]) -> list[str]:' benchmark/auto-resolve/scripts/frozen-verify-gate.py \
+  || ! grep -Fq 'path_has_actionable_solo_headroom_hypothesis(fixtures_root / fixture_id / "spec.md")' benchmark/auto-resolve/scripts/frozen-verify-gate.py \
+  || ! grep -Fq 'pair_trigger missing spec.solo_headroom_hypothesis' benchmark/auto-resolve/scripts/frozen-verify-gate.py \
+  || ! grep -Fq -- '--require-hypothesis-trigger' benchmark/auto-resolve/scripts/frozen-verify-gate.py \
+  || ! grep -Fq -- '--require-hypothesis-trigger' benchmark/auto-resolve/scripts/run-swebench-frozen-corpus.sh \
+  || ! grep -Fq 'missing-hypothesis-trigger' benchmark/auto-resolve/scripts/test-frozen-verify-gate.sh \
+  || ! grep -Fq '"pair_trigger_has_canonical_reason": has_canonical_pair_trigger_reason(trigger_reasons)' benchmark/auto-resolve/scripts/frozen-verify-gate.py \
+  || ! grep -Fq '"pair_trigger_has_canonical_reason": true' benchmark/auto-resolve/scripts/test-frozen-verify-gate.sh \
+  || ! grep -Fq '| Run | Fixture | Solo VERIFY | Pair VERIFY | Pair mode | Triggers | Wall ratio | External lift | Internal lift | Status | Reason |' benchmark/auto-resolve/scripts/frozen-verify-gate.py \
+  || ! grep -Fq '| Run | Fixture | Solo VERIFY | Pair VERIFY | Pair mode | Triggers | Wall ratio | External lift | Internal lift | Status | Reason |' benchmark/auto-resolve/scripts/test-frozen-verify-gate.sh \
+  || ! grep -Fq 'pair_trigger reasons missing known trigger reason' benchmark/auto-resolve/scripts/frozen-verify-gate.py \
+  || ! grep -Fq 'pair_trigger reasons contain unknown trigger reason' benchmark/auto-resolve/scripts/frozen-verify-gate.py \
+  || ! grep -Fq 'pair_trigger reasons missing canonical trigger reason' benchmark/auto-resolve/scripts/frozen-verify-gate.py \
+  || ! grep -Fq 'normalized-canonical-pair-trigger' benchmark/auto-resolve/scripts/test-frozen-verify-gate.sh \
+  || ! grep -Fq '| Fixture | Solo VERIFY | Pair VERIFY | Pair mode | Pair trigger | Triggers | Wall ratio | External lift | Internal lift | Included | Classification |' benchmark/auto-resolve/scripts/swebench-frozen-matrix.py \
+  || ! grep -Fq '| Fixture | Solo VERIFY | Pair VERIFY | Pair mode | Pair trigger | Triggers | Wall ratio | External lift | Internal lift | Included | Classification |' benchmark/auto-resolve/scripts/test-swebench-frozen-case.sh \
+  || ! grep -Fq 'def is_true(value: Any) -> bool:' benchmark/auto-resolve/scripts/swebench-frozen-matrix.py \
+  || ! grep -Fq 'swebench-bool-elapsed-test' benchmark/auto-resolve/scripts/test-swebench-frozen-case.sh \
+  || ! grep -Fq 'runner-nan-metadata' benchmark/auto-resolve/scripts/test-swebench-frozen-case.sh \
+  || ! grep -Fq 'runner-nan-expected' benchmark/auto-resolve/scripts/test-swebench-frozen-case.sh \
+  || ! grep -Fq 'json.loads(line, parse_constant=reject_json_constant)' benchmark/auto-resolve/scripts/run-swebench-solver-batch.sh \
+  || ! grep -Fq 'nan-instance-row' benchmark/auto-resolve/scripts/test-run-swebench-solver-batch.sh \
+  || ! grep -Fq 'manifest = loads_strict_json_object(pathlib.Path(sys.argv[1]).read_text())' benchmark/auto-resolve/scripts/run-iter-0033c.sh \
+  || ! grep -Fq 'manifest fixtures_pair_eligible must be a string array' benchmark/auto-resolve/scripts/run-iter-0033c.sh \
+  || ! grep -Fq 'parsed = json.loads(ln, parse_constant=reject_json_constant)' benchmark/auto-resolve/scripts/iter-0033c-compare.py \
+  || ! grep -Fq 'data = loads_strict_json_object(path.read_text())' benchmark/auto-resolve/scripts/judge.sh \
+  || ! grep -Fq 'decoder = json.JSONDecoder(parse_constant=reject_json_constant)' benchmark/auto-resolve/scripts/judge.sh \
+  || ! grep -Fq 'judge = loads_strict_json_object(judge_path.read_text())' benchmark/auto-resolve/scripts/judge-opus-pass.sh \
+  || ! grep -Fq 'gpt = loads_strict_json_object(pathlib.Path(sys.argv[2]).read_text())' benchmark/auto-resolve/scripts/judge-opus-pass.sh \
+  || ! grep -Fq 'decoder = json.JSONDecoder(parse_constant=reject_json_constant)' benchmark/auto-resolve/scripts/judge-opus-pass.sh \
+  || ! grep -Fq 'g = loads_strict_json_object(g_f.read_text())' benchmark/auto-resolve/scripts/judge-opus-pass.sh \
+  || ! grep -Fq 'swebench-malformed-pair-judge-test' benchmark/auto-resolve/scripts/test-swebench-frozen-case.sh \
+  || ! grep -Fq 'swebench-malformed-pair-trigger-test' benchmark/auto-resolve/scripts/test-swebench-frozen-case.sh \
+  || ! grep -Fq 'swebench-unknown-pair-trigger-test' benchmark/auto-resolve/scripts/test-swebench-frozen-case.sh \
+  || ! grep -Fq 'swebench-normalized-pair-trigger-test' benchmark/auto-resolve/scripts/test-swebench-frozen-case.sh \
+  || ! grep -Fq 'swebench-mixed-unknown-pair-trigger-test' benchmark/auto-resolve/scripts/test-swebench-frozen-case.sh \
+  || ! grep -Fq '"classification": "failed attempt: pair trigger contract: pair_trigger missing or malformed"' benchmark/auto-resolve/scripts/test-swebench-frozen-case.sh \
+  || ! grep -Fq '"classification": "failed attempt: pair trigger contract: pair_trigger reasons missing known trigger reason"' benchmark/auto-resolve/scripts/test-swebench-frozen-case.sh \
+  || ! grep -Fq '"classification": "failed attempt: pair trigger contract: pair_trigger reasons contain unknown trigger reason"' benchmark/auto-resolve/scripts/test-swebench-frozen-case.sh \
+  || ! grep -Fq '"classification": "failed attempt: pair trigger contract: pair_trigger reasons missing canonical trigger reason"' benchmark/auto-resolve/scripts/test-swebench-frozen-case.sh \
+  || ! grep -Fq 'historical-only-pair-trigger' benchmark/auto-resolve/scripts/test-frozen-verify-gate.sh \
+  || ! grep -Fq 'HISTORICAL_ONLY_TRIGGER_RUN_ID' benchmark/auto-resolve/scripts/test-swebench-frozen-case.sh \
+  || ! grep -Fq '| Pair trigger |' benchmark/auto-resolve/scripts/test-swebench-frozen-case.sh \
+  || ! grep -Fq '"verify_findings_count": "2"' benchmark/auto-resolve/scripts/test-swebench-frozen-case.sh \
+  || ! grep -Fq '"pair_found_more_low_or_worse": false' benchmark/auto-resolve/scripts/test-swebench-frozen-case.sh \
+  || ! grep -Fq 'pair-trigger eligibility/contract failures' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'swebench-string-bool-matrix-test' benchmark/auto-resolve/scripts/test-swebench-frozen-case.sh \
+  || ! grep -Fq 'findings = raw_findings if isinstance(raw_findings, list) else []' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'oracle artifact malformed or unreadable' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'scope-tier-a-nan' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'scope-tier-b-nan' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'expected.json malformed: spec_output_files must be a string array' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'loads_strict_json_object(exp_path.read_text())' benchmark/auto-resolve/scripts/oracle-scope-tier-a.py \
+  || ! grep -Fq 'loads_strict_json_object(pathlib.Path(args.expected).read_text())' benchmark/auto-resolve/scripts/oracle-scope-tier-b.py \
+  || ! grep -Fq 'failed attempt: malformed compare' benchmark/auto-resolve/scripts/test-swebench-frozen-case.sh \
+  || ! grep -Fq 'swebench-nan-matrix-test' benchmark/auto-resolve/scripts/test-swebench-frozen-case.sh \
+  || ! grep -Fq 'parse_prepared_case' benchmark/auto-resolve/scripts/prepare-swebench-frozen-corpus.py \
+  || ! grep -Fq 'loads_strict_json_object(path.read_text(encoding="utf8"))' benchmark/auto-resolve/scripts/prepare-swebench-frozen-case.py \
+  || ! grep -Fq 'parse_constant=reject_json_constant' benchmark/auto-resolve/scripts/collect-swebench-predictions.py \
+  || ! grep -Fq 'parse_constant=reject_json_constant' benchmark/auto-resolve/scripts/prepare-swebench-frozen-corpus.py \
+  || ! grep -Fq 'loads_strict_json_object(stdout)' benchmark/auto-resolve/scripts/prepare-swebench-frozen-corpus.py \
+  || ! grep -Fq 'batch-bad-timeout' benchmark/auto-resolve/scripts/test-swebench-frozen-case.sh \
+  || ! grep -Fq 'prepare-bad-timeout' benchmark/auto-resolve/scripts/test-swebench-frozen-case.sh \
+  || ! grep -Fq 'prepare-nan-case' benchmark/auto-resolve/scripts/test-swebench-frozen-case.sh \
+  || ! grep -Fq 'missing patch.diff for instance ids' benchmark/auto-resolve/scripts/test-swebench-frozen-case.sh \
+  || ! grep -Fq 'no non-empty patches collected' benchmark/auto-resolve/scripts/test-swebench-frozen-case.sh \
+  || ! grep -Fq 'collect-nan-instances' benchmark/auto-resolve/scripts/test-swebench-frozen-case.sh \
+  || ! grep -Fq 'batch-nan-predictions' benchmark/auto-resolve/scripts/test-swebench-frozen-case.sh \
+  || ! grep -Fq 'invalid JSON numeric constant: NaN' benchmark/auto-resolve/scripts/test-swebench-frozen-case.sh \
+  || ! grep -Fq 'fetch-empty-limit' benchmark/auto-resolve/scripts/test-swebench-frozen-case.sh \
+  || ! grep -Fq 'batch-empty-limit' benchmark/auto-resolve/scripts/test-swebench-frozen-case.sh \
+  || ! grep -Fq 'no prediction instances selected' benchmark/auto-resolve/scripts/test-swebench-frozen-case.sh \
+  || ! grep -Fq 'manifest malformed: expected JSON object' benchmark/auto-resolve/scripts/test-swebench-frozen-case.sh \
+  || ! grep -Fq 'corpus-manifest-nan' benchmark/auto-resolve/scripts/test-swebench-frozen-case.sh \
+  || ! grep -Fq 'loads_strict_json_object(manifest_path.read_text())' benchmark/auto-resolve/scripts/run-swebench-frozen-corpus.sh \
+  || ! grep -Fq 'loads_strict_json_object(pathlib.Path(sys.argv[1]).read_text())' benchmark/auto-resolve/scripts/run-swebench-frozen-corpus.sh \
+  || ! grep -Fq 'manifest malformed: prepared must be a non-empty array' benchmark/auto-resolve/scripts/test-swebench-frozen-case.sh \
+  || ! grep -Fq 'prepared[1] expected JSON object' benchmark/auto-resolve/scripts/test-swebench-frozen-case.sh \
+  || ! grep -Fq 'run ids malformed: no run ids' benchmark/auto-resolve/scripts/test-swebench-frozen-case.sh \
+  || ! grep -Fq 'run ids malformed: line 2 is empty' benchmark/auto-resolve/scripts/test-swebench-frozen-case.sh \
+  || ! grep -Fq 'run ids malformed: line 1 has unsafe run id' benchmark/auto-resolve/scripts/test-swebench-frozen-case.sh \
+  || ! grep -Fq -- '--fixture must match [A-Za-z0-9_.-]+' benchmark/auto-resolve/scripts/test-swebench-frozen-case.sh \
+  || ! grep -Fq -- '--run-id must match [A-Za-z0-9_.-]+' benchmark/auto-resolve/scripts/test-swebench-frozen-case.sh \
+  || ! grep -Fq 'unsafe SWE-bench repo' benchmark/auto-resolve/scripts/test-swebench-frozen-case.sh \
+  || ! grep -Fq 'unsafe SWE-bench base_commit' benchmark/auto-resolve/scripts/test-swebench-frozen-case.sh \
+  || ! grep -Fq 'parse_constant=reject_json_constant' benchmark/auto-resolve/scripts/prepare-swebench-solver-worktree.py \
+  || ! grep -Fq 'parse_constant=reject_json_constant' benchmark/auto-resolve/scripts/fetch-swebench-instances.py \
+  || ! grep -Fq 'malformed fetched row 1: row must be object' benchmark/auto-resolve/scripts/test-swebench-frozen-case.sh \
+  || ! grep -Fq 'solver-nan' benchmark/auto-resolve/scripts/test-swebench-frozen-case.sh \
+  || ! grep -Fq 'malformed scores_by_arm must not provide arm scores' benchmark/auto-resolve/scripts/test-iter-0033c-l1-summary.sh \
+  || ! grep -Fq 'non-dict result.json must not expose result fields' benchmark/auto-resolve/scripts/test-iter-0033c-l1-summary.sh \
+  || ! grep -Fq 'NaN result numeric fields must not appear in L1 summary' benchmark/auto-resolve/scripts/test-iter-0033c-l1-summary.sh \
+  || ! grep -Fq 'score_for(judge' benchmark/auto-resolve/scripts/iter-0033c-l1-summary.py \
+  || ! grep -Fq 'from pair_evidence_contract import is_score, is_strict_number' benchmark/auto-resolve/scripts/iter-0033c-l1-summary.py \
+  || ! grep -Fq 'loads_strict_json_object(path.read_text(encoding="utf8"))' benchmark/auto-resolve/scripts/iter-0033c-l1-summary.py \
+  || ! grep -Fq 'return legacy if is_score(legacy) else None' benchmark/auto-resolve/scripts/iter-0033c-l1-summary.py \
+  || ! grep -Fq 'def strict_number' benchmark/auto-resolve/scripts/iter-0033c-l1-summary.py \
+  || ! grep -Fq 'out-of-range scores must not appear in L1 summary' benchmark/auto-resolve/scripts/test-iter-0033c-l1-summary.sh \
+  || ! grep -Fq 'boolean scores must not appear in L1 summary' benchmark/auto-resolve/scripts/test-iter-0033c-l1-summary.sh \
+  || ! grep -Fq 'boolean result numeric fields must not appear in L1 summary' benchmark/auto-resolve/scripts/test-iter-0033c-l1-summary.sh \
+  || ! grep -Fq 'lowercase c_score' benchmark/auto-resolve/scripts/test-iter-0033c-l1-summary.sh; then
+  offenders="${offenders}"$'\n'"benchmark gates must require judge _blind_mapping before accepting score evidence"
+fi
+if ! grep -Fq 'solo_claude beats bare' benchmark/auto-resolve/scripts/build-pair-eligible-manifest.py \
+  || ! grep -Fq 'mapped_score' benchmark/auto-resolve/scripts/build-pair-eligible-manifest.py \
+  || ! grep -Fq 'parse_constant=reject_json_constant' benchmark/auto-resolve/scripts/build-pair-eligible-manifest.py \
+  || ! grep -Fq 'c1-nan-score' benchmark/auto-resolve/scripts/test-build-pair-eligible-manifest.sh \
+  || ! grep -Fq 'dirty F5 L1<=L0 row must not be promoted' benchmark/auto-resolve/scripts/test-build-pair-eligible-manifest.sh \
+  || ! grep -Fq 'l1-rerun-summary must not override pre-registered C1 selection grounds' benchmark/auto-resolve/scripts/test-build-pair-eligible-manifest.sh \
+  || ! grep -Fq 'wrong mapping' benchmark/auto-resolve/scripts/test-build-pair-eligible-manifest.sh; then
+  offenders="${offenders}"$'\n'"build-pair-eligible-manifest.py must use arm-mapped clean scores for F9 and L1<=L0 promotion"
+fi
+if [ -d benchmark/auto-resolve/results/20260510-f16-f23-f25-combined-proof/F16-cli-quote-tax-rules ] \
+  && ! python3 benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py \
+    --run-id 20260510-f16-f23-f25-combined-proof \
+    --pair-arm l2_risk_probes \
+    --min-fixtures 3 \
+    --min-pair-margin 5 \
+    --max-pair-solo-wall-ratio 3 \
+    --out-json /tmp/devlyn-lint-f16-f23-f25-pair-gate.json \
+    --out-md /tmp/devlyn-lint-f16-f23-f25-pair-gate.md >/dev/null 2>&1; then
+  offenders="${offenders}"$'\n'"current local artifacts for 20260510-f16-f23-f25-combined-proof must re-gate as PASS"
+fi
+if [ -d benchmark/auto-resolve/results/20260511-f21-current-riskprobes-v1/F21-cli-scheduler-priority ] \
+  && ! python3 benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py \
+    --run-id 20260511-f21-current-riskprobes-v1 \
+    --pair-arm l2_risk_probes \
+    --min-fixtures 1 \
+    --min-pair-margin 5 \
+    --max-pair-solo-wall-ratio 3 \
+    --out-json /tmp/devlyn-lint-f21-pair-gate.json \
+    --out-md /tmp/devlyn-lint-f21-pair-gate.md >/dev/null 2>&1; then
+  offenders="${offenders}"$'\n'"current local artifacts for 20260511-f21-current-riskprobes-v1 must re-gate as PASS"
+fi
+if ! grep -Fq 'bash scripts/lint-fixtures.sh' benchmark/auto-resolve/README.md; then
+  offenders="${offenders}"$'\n'"benchmark/auto-resolve/README.md: gate-change instructions must include fixture schema lint"
+fi
+if ! python3 - <<'PY'
+import runpy
+
+runtime = runpy.run_path("config/skills/_shared/verify-merge-findings.py")
+benchmark = runpy.run_path("benchmark/auto-resolve/scripts/pair_evidence_contract.py")
+
+runtime_known = set(runtime["KNOWN_PAIR_TRIGGER_REASONS"])
+canonical = set(benchmark["CANONICAL_PAIR_TRIGGER_REASONS"])
+aliases = set(benchmark["HISTORICAL_PAIR_TRIGGER_REASON_ALIASES"])
+normalized_aliases = set(benchmark["HISTORICAL_NORMALIZED_PAIR_TRIGGER_REASON_ALIASES"])
+known = set(benchmark["KNOWN_PAIR_TRIGGER_REASONS"])
+expected_aliases = {"risk_profile.high_risk", "risk_probes_enabled"}
+expected_normalized_aliases = {
+    "complexity.high.spec.frontmatter",
+    "frontmatter.complexity.high",
+    "high.complexity.spec",
+    "high.risk.profile",
+    "spec.frontmatter.complexity.high",
+    "state.complexity.high",
+}
+
+errors = []
+if canonical != runtime_known:
+    errors.append("benchmark canonical pair-trigger reasons must match runtime reasons")
+if aliases != expected_aliases:
+    errors.append("benchmark historical pair-trigger aliases changed unexpectedly")
+if normalized_aliases != expected_normalized_aliases:
+    errors.append("benchmark normalized historical pair-trigger aliases changed unexpectedly")
+if known != canonical | aliases:
+    errors.append("benchmark known pair-trigger reasons must be canonical plus aliases")
+if runtime_known & aliases:
+    errors.append("runtime pair-trigger reasons must not accept benchmark-only aliases")
+if not benchmark["has_historical_pair_trigger_reason"](["risk_profile.high_risk"]):
+    errors.append("benchmark historical alias helper must detect exact aliases")
+if benchmark["has_historical_pair_trigger_reason"](["complexity.high"]):
+    errors.append("benchmark historical alias helper must not count canonical reasons")
+if benchmark["is_known_pair_trigger_reason"]("risk high"):
+    errors.append("benchmark pair-trigger reader must reject normalized canonical reason strings")
+if benchmark["is_canonical_pair_trigger_reason"]("risk high"):
+    errors.append("benchmark pair-trigger canonical helper must be exact-match only")
+if not benchmark["is_known_pair_trigger_reason"]("high.risk.profile"):
+    errors.append("benchmark pair-trigger reader must preserve documented normalized historical aliases")
+if benchmark["is_canonical_pair_trigger_reason"]("high.risk.profile"):
+    errors.append("benchmark normalized historical aliases must not count as canonical reasons")
+if errors:
+    raise SystemExit("\n".join(errors))
+PY
+then
+  offenders="${offenders}"$'\n'"benchmark pair-trigger aliases must stay benchmark-only and runtime canonical reasons must stay in sync"
+fi
+	if ! grep -Fq 'bash scripts/lint-shadow-fixtures.sh' benchmark/auto-resolve/README.md \
+	  || ! grep -Fq 'The headroom and pair candidate runners' benchmark/auto-resolve/README.md \
+	  || ! grep -Fq 'accept explicitly named `S*` ids for dry-run checks and candidate measurement' benchmark/auto-resolve/README.md \
+	  || ! grep -Fq 'Use `run-suite.sh --suite shadow` only with `--dry-run`' benchmark/auto-resolve/README.md \
+	  || ! grep -Fq 'rejected/smoke controls do not' benchmark/auto-resolve/README.md \
+	  || ! grep -Fq 'npx devlyn-cli benchmark headroom --dry-run --min-fixtures 1 S1-cli-lang-flag' benchmark/auto-resolve/shadow-fixtures/README.md \
+	  || ! grep -Fq 'Use non-dry-run headroom/pair only for' benchmark/auto-resolve/shadow-fixtures/README.md \
+	  || ! grep -Fq 'explicitly named `S*` candidates with a solo-headroom hypothesis' benchmark/auto-resolve/shadow-fixtures/README.md \
+	  || ! grep -Fq 'promote a validated `S*` task to an active `F*`' benchmark/auto-resolve/shadow-fixtures/README.md \
+  || ! grep -Fq 'shadow suite run-suite is dry-run only' benchmark/auto-resolve/scripts/run-suite.sh \
+  || ! grep -Fq 'run-suite.sh --suite shadow --dry-run' benchmark/auto-resolve/scripts/run-suite.sh \
+  || ! grep -Fq 'shadow suite refuses provider/judge runs' benchmark/auto-resolve/scripts/run-suite.sh \
+  || ! grep -Fq 'npx devlyn-cli benchmark suite --suite shadow --dry-run' bin/devlyn.js \
+  || ! grep -Fq 'Use benchmark headroom/pair with explicit S* candidates for real provider measurement.' benchmark/auto-resolve/scripts/run-suite.sh \
+  || ! grep -Fq 'shadow-suite-provider-run' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'shadow-suite-judge-only-provider-run' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'arg-parse-shadow-cli-suite-dry-run' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'shadow suite dry-run must not invite a blocked non-dry-run suite invocation' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+	  || ! grep -Fq '`run-suite.sh --suite shadow` is dry-run only' benchmark/auto-resolve/shadow-fixtures/README.md \
+	  || ! grep -Fq 'parallel to the active golden `F*` fixtures' benchmark/auto-resolve/shadow-fixtures/README.md \
+	  || ! grep -Fq 'Active golden `F*` fixtures and pair-evidence audits control' benchmark/auto-resolve/shadow-fixtures/README.md \
+	  || ! grep -Fq 'Default `--suite golden` keeps active golden `F*` behavior' benchmark/auto-resolve/shadow-fixtures/README.md \
+	  || ! grep -Fq 'before any bare/solo/pair measurement' benchmark/auto-resolve/shadow-fixtures/README.md \
+	  || ! grep -Fq 'Do not spend' benchmark/auto-resolve/shadow-fixtures/README.md \
+	  || ! grep -Fq 'real provider calls on S1' benchmark/auto-resolve/shadow-fixtures/README.md \
+	  || ! grep -Fq 'not a solo<pair evidence candidate' benchmark/auto-resolve/shadow-fixtures/README.md \
+	  || grep -Fq 'F1-F9 still controls release' benchmark/auto-resolve/shadow-fixtures/README.md \
+	  || grep -Fq 'existing F1-F9 behavior' benchmark/auto-resolve/shadow-fixtures/README.md \
+	  || grep -Fq 'L0/L1/L2 measurement' benchmark/auto-resolve/shadow-fixtures/README.md; then
+  offenders="${offenders}"$'\n'"shadow fixture docs must describe lint and S* headroom/pair candidate calibration before golden promotion"
+fi
+if ! grep -Fq 'write a solo-headroom hypothesis' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'each new pair-candidate shadow fixture needs a' benchmark/auto-resolve/shadow-fixtures/README.md \
+  || ! grep -Fq "candidate's \`spec.md\`: name the visible behavior a capable \`solo_claude\`" benchmark/auto-resolve/README.md \
+  || ! grep -Fq '`solo_claude` baseline is expected to miss' benchmark/auto-resolve/shadow-fixtures/README.md \
+  || ! grep -Fq "candidate's \`spec.md\`" benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'solo-headroom hypothesis in `spec.md`' benchmark/auto-resolve/shadow-fixtures/README.md \
+  || ! grep -Fq 'observable command from `expected.json`' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'observable command from' benchmark/auto-resolve/shadow-fixtures/README.md \
+  || ! grep -Fq 'candidate runners enforce this as an actionable hypothesis' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'runners enforce this as an actionable hypothesis' benchmark/auto-resolve/shadow-fixtures/README.md \
+  || ! grep -Fq 'backticked observable command matching `expected.json`' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'backticked observable command matching `expected.json`' benchmark/auto-resolve/shadow-fixtures/README.md \
+	  || ! grep -Fq 'command/observable' benchmark/auto-resolve/README.md \
+	  || ! grep -Fq 'command/observable' benchmark/auto-resolve/shadow-fixtures/README.md \
+	  || ! grep -Fq 'itself containing `miss`' benchmark/auto-resolve/README.md \
+	  || ! grep -Fq 'itself containing `miss`' benchmark/auto-resolve/shadow-fixtures/README.md \
+	  || ! grep -Fq '## Solo ceiling avoidance' benchmark/auto-resolve/README.md \
+	  || ! grep -Fq '## Solo ceiling avoidance' benchmark/auto-resolve/shadow-fixtures/README.md \
+	  || ! grep -Fq 'solo-saturated `S2`-`S6` controls' benchmark/auto-resolve/README.md \
+	  || ! grep -Fq 'calibrated solo-saturated controls (`S2`-`S6`)' benchmark/auto-resolve/shadow-fixtures/README.md; then
+	  offenders="${offenders}"$'\n'"benchmark shadow candidate docs must require solo-headroom hypothesis and solo-ceiling avoidance before provider spend"
+	fi
+if [ ! -f benchmark/auto-resolve/scripts/solo-headroom-hypothesis.py ] \
+  || ! grep -Fq 'def has_actionable_solo_headroom_hypothesis_text' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq '"solo-headroom hypothesis" in lower' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'and "solo_claude" in lower' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'and "miss" in lower' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'OBSERVABLE_COMMAND_MARKERS = ("command", "observable", "expose")' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'def is_command_like_backtick' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'def path_has_actionable_solo_headroom_hypothesis' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'has_actionable_solo_headroom_hypothesis_text(text)' benchmark/auto-resolve/scripts/solo-headroom-hypothesis.py \
+  || ! grep -Fq 'actionable_observable_commands(text)' benchmark/auto-resolve/scripts/solo-headroom-hypothesis.py \
+  || ! grep -Fq -- '--expected-json' benchmark/auto-resolve/scripts/solo-headroom-hypothesis.py \
+  || ! grep -Fq 'expected_commands(args.expected_json)' benchmark/auto-resolve/scripts/solo-headroom-hypothesis.py \
+  || ! grep -Fq 'expected UTF-8 text' benchmark/auto-resolve/scripts/solo-headroom-hypothesis.py \
+  || ! grep -Fq 'solo-headroom-hypothesis.py' scripts/lint-shadow-fixtures.sh \
+  || ! grep -Fq 'has_solo_ceiling_avoidance_note' scripts/lint-shadow-fixtures.sh \
+  || ! grep -Fq 'solo-ceiling-avoidance.py' scripts/lint-shadow-fixtures.sh \
+  || ! grep -Fq 'SECTION_RE' benchmark/auto-resolve/scripts/solo-ceiling-avoidance.py \
+  || ! grep -Fq 'CONTROL_RE' benchmark/auto-resolve/scripts/solo-ceiling-avoidance.py \
+  || ! grep -Fq 'REASON_RE' benchmark/auto-resolve/scripts/solo-ceiling-avoidance.py \
+  || ! grep -Fq 'expected UTF-8 text' benchmark/auto-resolve/scripts/solo-ceiling-avoidance.py \
+  || ! grep -Fq 'fixture_has_solo_ceiling_avoidance_note' benchmark/auto-resolve/scripts/run-headroom-candidate.sh \
+  || ! grep -Fq 'fixture_has_solo_ceiling_avoidance_note' benchmark/auto-resolve/scripts/run-full-pipeline-pair-candidate.sh \
+  || ! grep -Fq 'solo-ceiling-avoidance.py' benchmark/auto-resolve/scripts/run-headroom-candidate.sh \
+  || ! grep -Fq 'solo-ceiling-avoidance.py' benchmark/auto-resolve/scripts/run-full-pipeline-pair-candidate.sh \
+  || ! grep -Fq 'shadow fixture NOTES.md needs ## Solo ceiling avoidance' benchmark/auto-resolve/scripts/run-headroom-candidate.sh \
+  || ! grep -Fq 'shadow fixture NOTES.md needs ## Solo ceiling avoidance' benchmark/auto-resolve/scripts/run-full-pipeline-pair-candidate.sh \
+  || ! grep -Fq 'solo-headroom-hypothesis.py' benchmark/auto-resolve/scripts/run-headroom-candidate.sh \
+  || ! grep -Fq 'solo-headroom-hypothesis.py' benchmark/auto-resolve/scripts/run-full-pipeline-pair-candidate.sh \
+  || ! grep -Fq -- '--expected-json "$dir/expected.json" "$dir/spec.md"' benchmark/auto-resolve/scripts/run-headroom-candidate.sh \
+  || ! grep -Fq -- '--expected-json "$dir/expected.json" "$dir/spec.md"' benchmark/auto-resolve/scripts/run-full-pipeline-pair-candidate.sh \
+  || ! grep -Fq -- '--expected-json "$d/expected.json" "$spec"' scripts/lint-shadow-fixtures.sh \
+  || ! grep -Fq 'from expected.json before provider spend' benchmark/auto-resolve/scripts/test-run-headroom-candidate.sh \
+  || ! grep -Fq 'from expected.json before provider spend' benchmark/auto-resolve/scripts/test-run-full-pipeline-pair-candidate.sh \
+  || ! grep -Fq 'unrelated-backtick-solo-headroom-hypothesis' benchmark/auto-resolve/scripts/test-run-headroom-candidate.sh \
+  || ! grep -Fq 'unrelated-backtick-solo-headroom-hypothesis' benchmark/auto-resolve/scripts/test-run-full-pipeline-pair-candidate.sh \
+  || ! grep -Fq 'observable-without-miss-solo-headroom-hypothesis' benchmark/auto-resolve/scripts/test-run-headroom-candidate.sh \
+  || ! grep -Fq 'observable-without-miss-solo-headroom-hypothesis' benchmark/auto-resolve/scripts/test-run-full-pipeline-pair-candidate.sh \
+  || ! grep -Fq 'missing-solo-ceiling-avoidance' benchmark/auto-resolve/scripts/test-run-headroom-candidate.sh \
+  || ! grep -Fq 'missing-solo-ceiling-avoidance' benchmark/auto-resolve/scripts/test-run-full-pipeline-pair-candidate.sh \
+  || ! grep -Fq 'weak-solo-ceiling-avoidance' benchmark/auto-resolve/scripts/test-run-headroom-candidate.sh \
+  || ! grep -Fq 'weak-solo-ceiling-avoidance' benchmark/auto-resolve/scripts/test-run-full-pipeline-pair-candidate.sh \
+  || ! grep -Fq 'hypothesis with unrelated backtick must fail' benchmark/auto-resolve/scripts/test-lint-fixtures.sh \
+  || ! grep -Fq 'weak hypothesis without observable command must fail' benchmark/auto-resolve/scripts/test-lint-fixtures.sh \
+  || ! grep -Fq 'actionable-hypothesis.md' benchmark/auto-resolve/scripts/test-lint-fixtures.sh \
+  || ! grep -Fq 'docs-style-actionable-hypothesis.md' benchmark/auto-resolve/scripts/test-lint-fixtures.sh \
+  || ! grep -Fq 'shadow-missing-solo-ceiling-avoidance' benchmark/auto-resolve/scripts/test-lint-fixtures.sh \
+  || ! grep -Fq 'shadow-weak-solo-ceiling-avoidance' benchmark/auto-resolve/scripts/test-lint-fixtures.sh \
+  || ! grep -Fq 'weak solo ceiling avoidance must fail' benchmark/auto-resolve/scripts/test-lint-fixtures.sh \
+  || ! grep -Fq 'non-utf8-solo-ceiling.md' benchmark/auto-resolve/scripts/test-lint-fixtures.sh \
+  || ! grep -Fq 'non-utf8-hypothesis.md' benchmark/auto-resolve/scripts/test-lint-fixtures.sh; then
+  offenders="${offenders}"$'\n'"solo-headroom hypothesis provider-spend guards must share one actionable checker and test weak-vs-actionable cases"
+fi
+if ! grep -Fq 'pair-candidate-frontier.py' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'npx devlyn-cli benchmark frontier --out-md /tmp/devlyn-pair-frontier.md' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'npx devlyn-cli benchmark frontier --out-md /tmp/devlyn-pair-frontier.md' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'npx devlyn-cli benchmark audit --out-dir /tmp/devlyn-benchmark-audit' README.md \
+  || ! grep -Fq 'npx devlyn-cli benchmark audit --out-dir /tmp/devlyn-benchmark-audit' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'npx devlyn-cli benchmark audit --out-dir /tmp/devlyn-benchmark-audit' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'npx devlyn-cli benchmark audit --require-hypothesis-trigger --out-dir /tmp/devlyn-benchmark-audit-strict' README.md \
+  || ! grep -Fq 'npx devlyn-cli benchmark audit --require-hypothesis-trigger --out-dir /tmp/devlyn-benchmark-audit-strict' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'npx devlyn-cli benchmark audit --require-hypothesis-trigger --out-dir /tmp/devlyn-benchmark-audit-strict' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'npx devlyn-cli benchmark audit --require-hypothesis-trigger --out-dir /tmp/devlyn-benchmark-audit-strict' bin/devlyn.js \
+  || ! grep -Fq 'npx devlyn-cli benchmark audit --require-hypothesis-trigger --out-dir /tmp/devlyn-benchmark-audit-strict' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'npx devlyn-cli benchmark frontier            Show pair candidate frontier scores/triggers without providers' bin/devlyn.js \
+  || ! grep -Fq 'npx devlyn-cli benchmark frontier            Show pair candidate frontier scores/triggers without providers' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'npx devlyn-cli benchmark audit               Audit pair evidence readiness' bin/devlyn.js \
+  || ! grep -Fq 'npx devlyn-cli benchmark audit               Audit pair evidence readiness' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'Show active rejected/evidence/unmeasured pair candidates, scores, and triggers without providers' bin/devlyn.js \
+  || ! grep -Fq 'Show active rejected/evidence/unmeasured pair candidates, scores, and triggers without providers' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'Prints pair evidence score rows with trigger reasons; --out-md includes a Triggers column' bin/devlyn.js \
+  || ! grep -Fq 'Prints pair evidence score rows with trigger reasons; --out-md includes a Triggers column' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'Prints frontier score rows plus headroom and pair quality handoff rows' bin/devlyn.js \
+  || ! grep -Fq 'Prints frontier score rows plus headroom and pair quality handoff rows' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'Prints frontier score rows plus headroom_rejections=PASS/FAIL, pair_evidence_quality=PASS/FAIL, pair_trigger_reasons=PASS/FAIL, pair_evidence_hypotheses=PASS/FAIL, pair_evidence_hypothesis_triggers=PASS/WARN/FAIL, historical-alias, and hypothesis-trigger gap handoff rows' bin/devlyn.js \
+  || ! grep -Fq 'Prints frontier score rows plus headroom_rejections=PASS/FAIL, pair_evidence_quality=PASS/FAIL, pair_trigger_reasons=PASS/FAIL, pair_evidence_hypotheses=PASS/FAIL, pair_evidence_hypothesis_triggers=PASS/WARN/FAIL, historical-alias, and hypothesis-trigger gap handoff rows' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq "audit: 'audit-pair-evidence.py'" bin/devlyn.js \
+  || ! grep -Fq 'npx devlyn-cli benchmark audit [options]' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq -- '--min-pair-evidence N  default: 4' bin/devlyn.js \
+  || ! grep -Fq -- '--min-pair-evidence N  default: 4' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq -- '--min-pair-margin N  default: 5' bin/devlyn.js \
+  || ! grep -Fq -- '--min-pair-margin N  default: 5' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq -- '--max-pair-solo-wall-ratio N  default: 3' bin/devlyn.js \
+  || ! grep -Fq -- '--max-pair-solo-wall-ratio N  default: 3' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq -- '--require-hypothesis-trigger' bin/devlyn.js \
+  || ! grep -Fq -- '--require-hypothesis-trigger' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq -- '--require-hypothesis-trigger' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'npx devlyn-cli benchmark audit-headroom --out-json /tmp/devlyn-headroom-audit.json' README.md \
+  || ! grep -Fq 'npx devlyn-cli benchmark audit-headroom --out-json /tmp/devlyn-headroom-audit.json' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'npx devlyn-cli benchmark audit-headroom --out-json /tmp/devlyn-headroom-audit.json' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'npx devlyn-cli benchmark audit-headroom      Audit failed headroom results' bin/devlyn.js \
+  || ! grep -Fq 'npx devlyn-cli benchmark audit-headroom      Audit failed headroom results' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq "'audit-headroom': 'audit-headroom-rejections.py'" bin/devlyn.js \
+  || ! grep -Fq 'npx devlyn-cli benchmark audit-headroom [options]' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'benchmarkMode === '\''frontier'\''' bin/devlyn.js \
+  || ! grep -Fq 'test-pair-candidate-frontier.sh' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'audit-pair-evidence.py' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'PASS audit-pair-evidence' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'FAIL audit-pair-evidence' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'FAIL audit-pair-evidence' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'audit.json' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'audit.json' README.md \
+  || ! grep -Fq 'audit.json' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'audit.json' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq '"frontier_summary": frontier_summary' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq '"pair_evidence_rows": pair_evidence_rows' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'def load_pair_evidence_rows' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'def load_frontier_stdout_metrics' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq '"summary_rows": summary_rows' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq '"aggregate_rows": aggregate_rows' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq '"trigger_rows": trigger_rows' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq '"trigger_rows_match_count": trigger_rows == expected_rows' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq '"hypothesis_trigger_rows_match_count": hypothesis_trigger_rows == expected_rows' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'from pair_evidence_contract import' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'from pair_evidence_contract import' benchmark/auto-resolve/scripts/audit-headroom-rejections.py \
+  || ! grep -Fq 'from pair_evidence_contract import' benchmark/auto-resolve/scripts/pair-candidate-frontier.py \
+  || ! grep -Fq 'def reject_json_constant' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'def loads_strict_json_object' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'parse_constant=reject_json_constant' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'loads_strict_json_object(path.read_text())' benchmark/auto-resolve/scripts/pair-candidate-frontier.py \
+  || ! grep -Fq 'loads_strict_json_object(path.read_text())' benchmark/auto-resolve/scripts/headroom-gate.py \
+  || ! grep -Fq 'loads_strict_json_object(path.read_text())' benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py \
+  || ! grep -Fq 'loads_strict_json_object(path.read_text())' benchmark/auto-resolve/scripts/audit-headroom-rejections.py \
+  || ! grep -Fq 'loads_strict_json_object(path.read_text())' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'nan-json-constant' benchmark/auto-resolve/scripts/test-pair-candidate-frontier.sh \
+  || ! grep -Fq 'nan-result-json' benchmark/auto-resolve/scripts/test-headroom-gate.sh \
+  || ! grep -Fq 'nan-result-json' benchmark/auto-resolve/scripts/test-full-pipeline-pair-gate.sh \
+  || ! grep -Fq 'CANONICAL_PAIR_TRIGGER_REASONS = {' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'HISTORICAL_PAIR_TRIGGER_REASON_ALIASES = {' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'HISTORICAL_NORMALIZED_PAIR_TRIGGER_REASON_ALIASES = {' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'Benchmark readers accept historical aliases only for archived artifacts' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'KNOWN_PAIR_TRIGGER_REASONS = (' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'mode.pair-verify' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'def normalized_pair_trigger_reason' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'def is_canonical_pair_trigger_reason' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'def has_known_pair_trigger_reason' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'def all_known_pair_trigger_reasons' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'def has_canonical_pair_trigger_reason' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'spec.solo_headroom_hypothesis' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'risk_profile.high_risk' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'has_known_pair_trigger_reason(reasons)' benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py \
+  || ! grep -Fq 'all_known_pair_trigger_reasons(reasons)' benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py \
+  || ! grep -Fq 'has_canonical_pair_trigger_reason(reasons)' benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py \
+  || ! grep -Fq '"pair_trigger_has_canonical_reason": has_canonical_pair_trigger_reason(' benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py \
+  || ! grep -Fq '"pair_trigger_has_hypothesis_reason": (' benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py \
+  || ! grep -Fq '"require_hypothesis_trigger": args.require_hypothesis_trigger' benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py \
+  || ! grep -Fq '"pair_trigger_has_canonical_reason": true' benchmark/auto-resolve/scripts/test-full-pipeline-pair-gate.sh \
+  || ! grep -Fq '"pair_trigger_has_hypothesis_reason": true' benchmark/auto-resolve/scripts/test-full-pipeline-pair-gate.sh \
+  || ! grep -Fq '"require_hypothesis_trigger": true' benchmark/auto-resolve/scripts/test-full-pipeline-pair-gate.sh \
+  || ! grep -Fq 'pair_trigger eligible with a canonical reason' benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py \
+  || ! grep -Fq 'pair_trigger eligible with canonical reason' benchmark/auto-resolve/scripts/test-full-pipeline-pair-gate.sh \
+  || ! grep -Fq 'pair_trigger reasons missing known trigger reason' benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py \
+  || ! grep -Fq 'pair_trigger reasons contain unknown trigger reason' benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py \
+  || ! grep -Fq 'pair_trigger reasons missing canonical trigger reason' benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py \
+  || ! grep -Fq 'fixture_spec_has_solo_headroom_hypothesis(fixture_dir.name)' benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py \
+  || ! grep -Fq 'pair_trigger missing spec.solo_headroom_hypothesis' benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py \
+  || ! grep -Fq -- '--require-hypothesis-trigger' benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py \
+  || ! grep -Fq -- '--require-hypothesis-trigger' benchmark/auto-resolve/scripts/run-full-pipeline-pair-candidate.sh \
+  || ! grep -Fq 'release audit: npx devlyn-cli benchmark audit --require-hypothesis-trigger --out-dir /tmp/devlyn-benchmark-audit-strict' benchmark/auto-resolve/scripts/run-full-pipeline-pair-candidate.sh \
+  || ! grep -Fq 'release audit: npx devlyn-cli benchmark audit --require-hypothesis-trigger --out-dir /tmp/devlyn-benchmark-audit-strict' benchmark/auto-resolve/scripts/test-run-full-pipeline-pair-candidate.sh \
+  || ! grep -Fq 'missing-hypothesis-trigger' benchmark/auto-resolve/scripts/test-full-pipeline-pair-gate.sh \
+  || ! grep -Fq 'grep -Fq -- '\''--require-hypothesis-trigger'\''' benchmark/auto-resolve/scripts/test-run-full-pipeline-pair-candidate.sh \
+  || ! grep -Fq 'def fmt_trigger_reasons' benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py \
+  || ! grep -Fq 'Hypothesis trigger required: {str(report' benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py \
+  || ! grep -Fq '| Fixture | Bare | Bare headroom | Solo_claude | Solo_claude headroom | Pair | Margin | Pair mode | Hypothesis trigger | Triggers | Wall ratio | Status | Reason |' benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py \
+  || ! grep -Fq '| Fixture | Bare | Bare headroom | Solo_claude | Solo_claude headroom | Pair | Margin | Pair mode | Hypothesis trigger | Triggers | Wall ratio | Status | Reason |' benchmark/auto-resolve/scripts/test-full-pipeline-pair-gate.sh \
+  || ! grep -Fq 'trigger-reason, and wall-ratio columns' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'eligible with non-empty reasons and at least one canonical reason' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'fixtures with an actionable solo-headroom hypothesis must include `spec.solo_headroom_hypothesis` in the trigger reasons' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'pair trigger eligibility, trigger reasons, canonical-trigger coverage, and `spec.solo_headroom_hypothesis` coverage' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'unknown-pair-trigger-reason' benchmark/auto-resolve/scripts/test-full-pipeline-pair-gate.sh \
+  || ! grep -Fq 'mixed-unknown-pair-trigger-reason' benchmark/auto-resolve/scripts/test-full-pipeline-pair-gate.sh \
+  || ! grep -Fq 'normalized-canonical-pair-trigger-reason' benchmark/auto-resolve/scripts/test-full-pipeline-pair-gate.sh \
+  || ! grep -Fq 'historical-only-pair-trigger-reason' benchmark/auto-resolve/scripts/test-full-pipeline-pair-gate.sh \
+  || ! grep -Fq 'unknown-pair-trigger' benchmark/auto-resolve/scripts/test-frozen-verify-gate.sh \
+  || ! grep -Fq 'def normalize_pair_evidence_row' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'def best_pair_evidence' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'ALLOWED_PAIR_ARMS = {"l2_risk_probes", "l2_gated"}' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'pair_arm not in ALLOWED_PAIR_ARMS' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'pair_mode = row.get("pair_mode")' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'if pair_mode is not True:' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'pair_trigger_eligible = row.get("pair_trigger_eligible")' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'if pair_trigger_eligible is not True:' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'pair_trigger_reasons = row.get("pair_trigger_reasons")' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'isinstance(pair_trigger_reasons, list)' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'all_known_pair_trigger_reasons(pair_trigger_reasons)' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'has_canonical_pair_trigger_reason(pair_trigger_reasons)' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq '"pair_trigger_reasons": pair_trigger_reasons' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq '"pair_trigger_has_canonical_reason": True' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq '"pair_trigger_has_hypothesis_reason": (' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'def pair_result_trigger_reasons' benchmark/auto-resolve/scripts/pair-candidate-frontier.py \
+  || ! grep -Fq 'candidate_row = dict(row)' benchmark/auto-resolve/scripts/pair-candidate-frontier.py \
+  || ! grep -Fq 'all_known_pair_trigger_reasons(reasons)' benchmark/auto-resolve/scripts/pair-candidate-frontier.py \
+  || ! grep -Fq 'has_canonical_pair_trigger_reason(reasons)' benchmark/auto-resolve/scripts/pair-candidate-frontier.py \
+  || ! grep -Fq 'candidate_row["pair_trigger_reasons"] = reasons' benchmark/auto-resolve/scripts/pair-candidate-frontier.py \
+  || ! grep -Fq '"reasons": ["complexity.high", "looks-hard"]' benchmark/auto-resolve/scripts/test-pair-candidate-frontier.sh \
+  || ! grep -Fq '"reasons":["risk high"]' benchmark/auto-resolve/scripts/test-pair-candidate-frontier.sh \
+  || ! grep -Fq 'triggers={triggers}' benchmark/auto-resolve/scripts/pair-candidate-frontier.py \
+  || ! grep -Fq 'hypothesis_trigger={hypothesis_trigger}' benchmark/auto-resolve/scripts/pair-candidate-frontier.py \
+  || ! grep -Fq '| Fixture | Status | Verdict | Evidence | Pair arm | Triggers | Hypothesis trigger |' benchmark/auto-resolve/scripts/pair-candidate-frontier.py \
+  || ! grep -Fq 'pair-arm, and trigger-reason columns' benchmark/auto-resolve/README.md \
+  || ! grep -Fq '| Fixture | Status | Verdict | Evidence | Pair arm | Triggers | Hypothesis trigger |' benchmark/auto-resolve/scripts/test-pair-candidate-frontier.sh \
+  || ! grep -Fq 'def check_pair_evidence_hypotheses' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'def pair_evidence_hypothesis_trigger_rows' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'def pair_evidence_hypothesis_trigger_gap_details' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'def check_pair_evidence_hypothesis_triggers' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'def print_pair_evidence_hypothesis_triggers_summary' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'pair_evidence_hypotheses={status}' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'pair_evidence_hypothesis_triggers={status}' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'pair evidence hypotheses missing for fixture(s)' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'pair-evidence-hypotheses' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'pair_evidence_hypotheses=PASS documented=2 total=2' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'pair_evidence_hypothesis_triggers=WARN matched=0 documented=2 total=2' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'pair_evidence_hypothesis_trigger_gaps=F16-cli-quote-tax-rules=complexity.high;F21-cli-scheduler-priority=complexity.high,risk_profile.high_risk' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'pair_evidence_hypothesis_triggers=FAIL matched=0 documented=2 total=2' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'strict benchmark audit must not report current hypothesis-trigger gaps' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'pair_evidence_hypothesis_triggers=PASS matched=4 documented=4 total=4' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'verdict=pair_evidence_passed triggers={triggers}' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'def format_trigger_reasons' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'missing-frontier-triggers' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'triggers=complexity.high' benchmark/auto-resolve/scripts/test-pair-candidate-frontier.sh \
+  || ! grep -Fq '"pair_trigger_reasons": pair_trigger_reasons(pair_result)' benchmark/auto-resolve/scripts/full-pipeline-pair-gate.py \
+  || ! grep -Fq '"pair_trigger_eligible": pair_trigger_eligible' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq '"verdict": "pair_evidence_passed"' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq '"pair_mode": pair_mode' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'def pair_result_trigger_reasons' benchmark/auto-resolve/scripts/audit-headroom-rejections.py \
+  || ! grep -Fq 'candidate_row = dict(row)' benchmark/auto-resolve/scripts/audit-headroom-rejections.py \
+  || ! grep -Fq 'candidate_row["pair_trigger_reasons"] = reasons' benchmark/auto-resolve/scripts/audit-headroom-rejections.py \
+  || ! grep -Fq 'F16-cli-quote-tax-rules/l2_risk_probes/result.json' benchmark/auto-resolve/scripts/test-audit-headroom-rejections.sh \
+  || ! grep -Fq '"reasons":["risk high"]' benchmark/auto-resolve/scripts/test-audit-headroom-rejections.sh \
+  || ! grep -Fq 'def is_strict_int' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'def is_score' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq '0 <= value <= 100' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'def is_strict_number' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'import math' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'math.isfinite(value)' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'and value > 0' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'if pair_margin != pair_score - solo_score:' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq '"min_pair_evidence": min_pair_evidence' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq '"actual_rows": len(pair_evidence_rows)' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq '"rows_match_count": (' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq '"pair_evidence_quality": {' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'def check_pair_evidence_quality' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'def print_pair_evidence_quality' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'def check_pair_trigger_reasons' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'def print_pair_trigger_reasons_summary' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'def pair_trigger_historical_alias_details' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'pair_evidence_quality={status} min_pair_margin_actual={actual_margin}' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'pair_trigger_reasons={status} canonical={canonical} historical_alias={historical_alias}' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq '"historical_alias_details": historical_alias_details' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'frontier pair_margin_min does not match pair evidence rows' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'frontier pair_solo_wall_ratio_max does not match pair evidence rows' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'is_strict_int(pair_evidence_count)' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'if not is_strict_int(count):' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'pair evidence count missing or malformed from frontier report' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'pair evidence rows {len(rows)} do not match summary count {count}' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'def check_frontier_report' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'frontier verdict {verdict!r} is not PASS' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'frontier has {unmeasured_count} unmeasured candidate fixture(s)' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq '"frontier_report": {' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq '"min_pair_margin": min_pair_margin' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq '"max_pair_solo_wall_ratio": max_pair_solo_wall_ratio' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'not math.isfinite(args.max_pair_solo_wall_ratio)' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'max-pair-solo-wall-ratio must be finite and > 0' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'pair evidence count {count} below required minimum {minimum}' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'assert audit["checks"]["min_pair_evidence"]["actual_rows"] == len(audit["pair_evidence_rows"])' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'assert audit["checks"]["frontier_report"]["verdict"] == frontier["verdict"]' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'pair_evidence_quality=PASS min_pair_margin_actual=+21 min_pair_margin_required=+5 max_wall_actual=2.25x max_wall_allowed=3.00x' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'pair_trigger_reasons=PASS canonical=4 historical_alias=0 exposed=4 total=4 summary=4 rows_match=true' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'current benchmark audit must not report historical aliases or hypothesis-trigger gaps' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'benchmark audit must fail when min pair evidence exceeds current evidence rows' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'actual_pair_evidence=$(python3 - "$TMP/audit/audit.json"' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'required_pair_evidence=$((actual_pair_evidence + 1))' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'pair evidence count ${actual_pair_evidence} below required minimum ${required_pair_evidence}' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'pair_margin_avg=+27.25 pair_margin_min=+21 wall_avg=1.66x wall_max=2.25x' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'F16-cli-quote-tax-rules: bare=50 solo_claude=75 pair=96 arm=l2_risk_probes margin=+21' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'FAIL audit-pair-evidence' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'assert audit["checks"]["min_pair_evidence"]["required"] == required' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'assert audit["checks"]["pair_evidence_quality"]["status"] == "PASS"' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'assert audit["checks"]["pair_trigger_reasons"]["summary_pair_evidence_count"] == 4' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'assert audit["checks"]["pair_trigger_reasons"]["rows_match_count"] is True' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'assert audit["checks"]["pair_trigger_reasons"]["canonical_rows"] == 4' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'assert audit["checks"]["pair_trigger_reasons"]["historical_alias_details"] == []' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'assert len(audit["pair_evidence_rows"]) == frontier["pair_evidence_count"]' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'benchmark frontier must fail when active unmeasured candidates remain' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'PASS pair-candidate-frontier' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'FAIL pair-candidate-frontier' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'assert report["rows"][0]["status"] == "candidate_unmeasured"' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'assert row["verdict"] == "pair_evidence_passed"' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'assert row["pair_mode"] is True' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'assert row["pair_trigger_eligible"] is True' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'assert isinstance(row["pair_solo_wall_ratio"], (int, float))' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'assert report["frontier_summary"]["pair_margin_avg"] == 27' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'assert report["pair_evidence_rows"] == [' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'assert report["checks"]["min_pair_evidence"]["actual_rows"] == 2' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'assert report["checks"]["min_pair_evidence"]["rows_match_count"] is True' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'assert report["checks"]["min_pair_evidence"]["status"] == "FAIL"' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'assert report["checks"]["pair_evidence_quality"]["status"] == "PASS"' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'assert report["checks"]["pair_trigger_reasons"]["status"] == "PASS"' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'assert report["checks"]["pair_trigger_reasons"]["summary_pair_evidence_count"] == 2' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'assert report["checks"]["pair_trigger_reasons"]["rows_match_count"] is True' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'assert report["checks"]["pair_trigger_reasons"]["historical_alias_details"] == [' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'normalized-canonical-trigger-reason-rows' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'assert report["checks"]["pair_evidence_quality"]["min_pair_margin_actual"] == 21' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'pair_evidence_quality=PASS min_pair_margin_actual=+21 min_pair_margin_required=+5 max_wall_actual=1.47x max_wall_allowed=3.00x' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'pair_evidence_quality=FAIL min_pair_margin_actual=+4 min_pair_margin_required=+5 max_wall_actual=1.20x max_wall_allowed=3.00x' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'no-quality-rows-frontier.json' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'pair evidence quality check has no complete rows' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'low-quality-frontier.json' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'pair evidence margin below minimum for fixture(s): F16-cli-quote-tax-rules' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'high-wall-frontier.json' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'pair evidence wall ratio above maximum for fixture(s): F16-cli-quote-tax-rules' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'summary-mismatch-frontier.json' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'frontier pair_margin_min does not match pair evidence rows' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'summary-wall-mismatch-frontier.json' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'frontier pair_solo_wall_ratio_max does not match pair evidence rows' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'frontier-fail-verdict.json' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'frontier-unmeasured.json' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'frontier-malformed-unmeasured.json' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'frontier-incomplete-best.json' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq '"run_id": "lower-complete"' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'bool-frontier-count.json' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'malformed-pair-evidence-count' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'mismatched-frontier-rows.json' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'pair evidence rows 1 do not match summary count 2' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'bad-frontier-row-fields.json' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'nan-frontier-row-fields.json' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'nan-max-wall-ratio' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'mismatched-margin-row-fields.json' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'mismatched-margin-row-fields' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'overrange-score-row-fields.json' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'overrange-score-row-fields' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'invalid-pair-arm-row-fields.json' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'invalid-pair-arm-row-fields' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'false-pair-mode-row-fields.json' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'false-pair-mode-row-fields' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'zero-wall-row-fields.json' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'zero-wall-row-fields' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'pair evidence count 2 below required minimum 4' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'assert rows["F22-cli-low-margin"]["status"] == "candidate_unmeasured"' benchmark/auto-resolve/scripts/test-pair-candidate-frontier.sh \
+  || ! grep -Fq 'assert rows["F22-cli-low-margin"]["status"] == "pair_evidence_passed"' benchmark/auto-resolve/scripts/test-pair-candidate-frontier.sh \
+  || ! grep -Fq 'assert rows["F23-cli-high-wall"]["status"] == "candidate_unmeasured"' benchmark/auto-resolve/scripts/test-pair-candidate-frontier.sh \
+  || ! grep -Fq 'assert rows["F23-cli-high-wall"]["status"] == "pair_evidence_passed"' benchmark/auto-resolve/scripts/test-pair-candidate-frontier.sh \
+  || ! grep -Fq 'def print_final_verdict' benchmark/auto-resolve/scripts/pair-candidate-frontier.py \
+  || ! grep -Fq 'PASS pair-candidate-frontier' benchmark/auto-resolve/scripts/pair-candidate-frontier.py \
+  || ! grep -Fq 'FAIL pair-candidate-frontier' benchmark/auto-resolve/scripts/pair-candidate-frontier.py \
+  || ! grep -Fq 'PASS pair-candidate-frontier' benchmark/auto-resolve/scripts/test-pair-candidate-frontier.sh \
+  || ! grep -Fq 'FAIL pair-candidate-frontier' benchmark/auto-resolve/scripts/test-pair-candidate-frontier.sh \
+  || ! grep -Fq 'requires at least four fixtures with passing pair evidence' README.md \
+  || ! grep -Fq 'revalidates frontier `verdict: PASS`, zero unmeasured candidates' README.md \
+  || ! grep -Fq 'requires at least four active fixtures with passing pair evidence' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'revalidates frontier `verdict: PASS`' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'requires at least four active fixtures with passing pair evidence' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'revalidates frontier `verdict: PASS`, zero unmeasured candidates' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'counted by `benchmark audit` as the fourth passing pair-evidence row' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'counted by `benchmark audit` as the fourth passing pair-evidence row' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'counted by `benchmark audit` as the fourth passing pair-evidence row' benchmark/auto-resolve/BENCHMARK-RESULTS.md \
+  || ! grep -Fq 'the default 5-point pair margin' README.md \
+  || ! grep -Fq -- '`--pair-verify` and `--no-pair` are mutually exclusive' README.md \
+  || ! grep -Fq 'default 5-point pair margin' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'the default 5-point pair margin' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq '3x pair/solo wall ratio' README.md \
+  || ! grep -Fq '3x pair/solo wall ratio' benchmark/auto-resolve/README.md \
+  || ! grep -Fq '3x pair/solo wall ratio' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'audit.json` with the frontier summary' README.md \
+  || ! grep -Fq 'audit.json` with the frontier summary' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'audit.json` with the frontier summary' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'artifact map' README.md \
+  || ! grep -Fq 'artifact map' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'artifact map' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq '`checks.frontier_stdout` records summary, aggregate, final-verdict, expected, printed score-row, trigger-visible row, and hypothesis-trigger-visible row counts' README.md \
+  || ! grep -Fq '`headroom_rejections=...`, `pair_evidence_quality=...`,' README.md \
+  || ! grep -Fq '`pair_trigger_reasons=...`, `pair_evidence_hypotheses=...`, and' README.md \
+  || ! grep -Fq '`pair_evidence_hypothesis_triggers=...` handoff rows' README.md \
+  || ! grep -Fq '`pair_trigger_historical_aliases=...` when archived evidence includes legacy' README.md \
+  || ! grep -Fq '`pair_evidence_hypothesis_trigger_gaps=...` when documented' README.md \
+  || ! grep -Fq 'canonical trigger reason coverage' README.md \
+  || ! grep -Fq '`checks.pair_evidence_quality` records the same quality thresholds from the compact rows' README.md \
+  || ! grep -Fq '`checks.pair_trigger_reasons` records canonical/historical-alias/exposed/total trigger-reason row counts, fixture-level historical alias details, summary count, and row-match status' README.md \
+  || ! grep -Fq '`checks.pair_evidence_hypotheses` records documented/total pair-evidence hypothesis row counts' README.md \
+  || ! grep -Fq '`checks.pair_evidence_hypothesis_triggers` records whether documented hypotheses also appear as `spec.solo_headroom_hypothesis` trigger reasons plus fixture-level gap details' README.md \
+  || ! grep -Fq 'regenerated pair evidence' README.md \
+  || ! grep -Fq 'Historical trigger aliases are only reported for archived artifact review' README.md \
+  || ! grep -Fq 'current pair-evidence gates fail historical-only or unknown trigger reasons' README.md \
+  || ! grep -Fq '`checks.frontier_stdout`' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'aggregate, final-verdict, expected, printed score-row, trigger-visible row, and hypothesis-trigger-visible row counts' benchmark/auto-resolve/README.md \
+  || ! grep -Fq '`headroom_rejections=...`,' benchmark/auto-resolve/README.md \
+  || ! grep -Fq '`pair_trigger_reasons=...`, and' benchmark/auto-resolve/README.md \
+  || ! grep -Fq '`pair_evidence_hypotheses=...` and' benchmark/auto-resolve/README.md \
+  || ! grep -Fq '`pair_evidence_hypothesis_triggers=...` handoff rows' benchmark/auto-resolve/README.md \
+  || ! grep -Fq '`pair_trigger_historical_aliases=...` when archived evidence includes legacy' benchmark/auto-resolve/README.md \
+  || ! grep -Fq '`pair_evidence_hypothesis_trigger_gaps=...` when documented' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'canonical trigger reason coverage' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'records the same quality thresholds from the compact rows,' benchmark/auto-resolve/README.md \
+  || ! grep -Fq '`checks.pair_trigger_reasons` records canonical/historical-alias/exposed/total trigger-reason row counts, fixture-level historical alias details, summary count, and row-match status' benchmark/auto-resolve/README.md \
+  || ! grep -Fq '`checks.pair_evidence_hypotheses` records documented/total pair-evidence hypothesis row counts' benchmark/auto-resolve/README.md \
+  || ! grep -Fq '`checks.pair_evidence_hypothesis_triggers` records whether documented hypotheses also appear as `spec.solo_headroom_hypothesis` trigger reasons plus fixture-level gap details' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'regenerated pair evidence' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'Historical trigger aliases are only reported for archived artifact review' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'current pair-evidence gates fail historical-only or unknown trigger reasons' benchmark/auto-resolve/README.md \
+  || ! grep -Fq '`checks.frontier_stdout` records summary, aggregate, final-verdict, expected, printed score-row, trigger-visible row, and hypothesis-trigger-visible row counts' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'also prints `headroom_rejections=...`, `pair_evidence_quality=...`,' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq '`pair_trigger_reasons=...`, `pair_evidence_hypotheses=...`, and `pair_evidence_hypothesis_triggers=...` handoff rows' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq '`pair_trigger_historical_aliases=...` when archived evidence includes legacy' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq '`pair_evidence_hypothesis_trigger_gaps=...` when documented' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'canonical trigger reason coverage' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq '`checks.pair_evidence_quality` records the same quality thresholds from the compact rows' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq '`checks.pair_trigger_reasons` records canonical/historical-alias/exposed/total trigger-reason row counts, fixture-level historical alias details, summary count, and row-match status' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq '`checks.pair_evidence_hypotheses` records documented/total pair-evidence hypothesis row counts' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq '`checks.pair_evidence_hypothesis_triggers` records whether documented hypotheses also appear as `spec.solo_headroom_hypothesis` trigger reasons plus fixture-level gap details' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'regenerated pair evidence' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'Historical trigger aliases are only reported for archived artifact review' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'current pair-evidence gates fail historical-only or unknown trigger reasons' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'compact trigger-backed verdict-bearing `pair_evidence_rows`' README.md \
+  || ! grep -Fq 'pair_trigger_eligible: true' README.md \
+  || ! grep -Fq 'non-empty `pair_trigger_reasons`, `pair_trigger_has_canonical_reason: true`, and `pair_trigger_has_hypothesis_reason`; the audit fails rows missing trigger reasons' README.md \
+  || ! grep -Fq 'pair_trigger_has_canonical_reason: true' README.md \
+  || ! grep -Fq '`pair_evidence_rows`' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'compact trigger-backed verdict-bearing score rows' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'pair_trigger_eligible: true' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'non-empty `pair_trigger_reasons`, `pair_trigger_has_canonical_reason: true`, and `pair_trigger_has_hypothesis_reason`; the audit fails rows missing trigger reasons' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'pair_trigger_has_canonical_reason: true' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'compact trigger-backed verdict-bearing `pair_evidence_rows`' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'pair_trigger_eligible: true' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'non-empty `pair_trigger_reasons`, `pair_trigger_has_canonical_reason: true`, and `pair_trigger_has_hypothesis_reason`; the audit fails rows missing trigger reasons' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'pair_trigger_has_canonical_reason: true' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'revalidates `pair_mode: true`' README.md \
+  || ! grep -Fq 'satisfy `pair_mode: true`' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'revalidates `pair_mode: true`' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'JSON rows expose `pair_trigger_reasons` and' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'Markdown output includes a `Triggers`' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'trigger reasons, canonical-trigger coverage, classification counts' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'Its Markdown table includes a `Triggers` column' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'compact evidence row count must match the' README.md \
+  || ! grep -Fq 'compact evidence row count must match the' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'evidence row count must match the frontier evidence count' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq '"verdict": "PASS" if unmeasured_candidate_total == 0 else "FAIL"' benchmark/auto-resolve/scripts/pair-candidate-frontier.py \
+  || ! grep -Fq 'assert frontier["unmeasured_count"] == 0' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'pair_margin_avg' benchmark/auto-resolve/scripts/pair-candidate-frontier.py \
+  || ! grep -Fq 'from pair_evidence_contract import' benchmark/auto-resolve/scripts/pair-candidate-frontier.py \
+  || ! grep -Fq 'from pair_evidence_contract import' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'from pair_evidence_contract import' benchmark/auto-resolve/scripts/audit-headroom-rejections.py \
+  || ! grep -Fq 'def normalize_pair_evidence_row' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'def best_pair_evidence' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'ALLOWED_PAIR_ARMS = {"l2_risk_probes", "l2_gated"}' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'pair_arm not in ALLOWED_PAIR_ARMS' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'pair_mode = row.get("pair_mode")' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'if pair_mode is not True:' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'pair_trigger_eligible = row.get("pair_trigger_eligible")' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'if pair_trigger_eligible is not True:' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq '"pair_trigger_eligible": pair_trigger_eligible' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq '"pair_mode": pair_mode' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'def is_strict_int' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'def is_score' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq '0 <= value <= 100' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'def is_strict_number' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'import math' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'math.isfinite(value)' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'and value > 0' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'if pair_margin != pair_score - solo_score:' benchmark/auto-resolve/scripts/pair_evidence_contract.py \
+  || ! grep -Fq 'not math.isfinite(args.max_pair_solo_wall_ratio)' benchmark/auto-resolve/scripts/pair-candidate-frontier.py \
+  || ! grep -Fq 'nan-wall-run' benchmark/auto-resolve/scripts/test-pair-candidate-frontier.sh \
+  || ! grep -Fq 'inflated-margin-run' benchmark/auto-resolve/scripts/test-pair-candidate-frontier.sh \
+  || ! grep -Fq 'overrange-score-run' benchmark/auto-resolve/scripts/test-pair-candidate-frontier.sh \
+  || ! grep -Fq 'invalid-arm-run' benchmark/auto-resolve/scripts/test-pair-candidate-frontier.sh \
+  || ! grep -Fq 'false-pair-mode-run' benchmark/auto-resolve/scripts/test-pair-candidate-frontier.sh \
+  || ! grep -Fq 'zero-wall-run' benchmark/auto-resolve/scripts/test-pair-candidate-frontier.sh \
+  || ! grep -Fq 'nan-max-wall-ratio' benchmark/auto-resolve/scripts/test-pair-candidate-frontier.sh \
+  || ! grep -Fq 'pair_margin_avg=+21.00 pair_margin_min=+21 wall_avg=1.28x wall_max=1.28x' benchmark/auto-resolve/scripts/test-pair-candidate-frontier.sh \
+  || ! grep -Fq 'incomplete-high-run' benchmark/auto-resolve/scripts/test-pair-candidate-frontier.sh \
+  || ! grep -Fq 'bad-pair-evidence-json' benchmark/auto-resolve/scripts/test-pair-candidate-frontier.sh \
+  || ! grep -Fq 'pair evidence artifact malformed' benchmark/auto-resolve/scripts/pair-candidate-frontier.py \
+  || ! grep -Fq 'bad-pair-evidence-rows' benchmark/auto-resolve/scripts/test-pair-candidate-frontier.sh \
+  || ! grep -Fq 'pair evidence artifact rows malformed' benchmark/auto-resolve/scripts/pair-candidate-frontier.py \
+  || ! grep -Fq 'assert len(rows["F16-cli-quote-tax-rules"]["passing_pair_evidence"]) == 1' benchmark/auto-resolve/scripts/test-pair-candidate-frontier.sh \
+  || ! grep -Fq 'plus row-level verdicts' README.md \
+  || ! grep -Fq 'including pair arm, trigger reasons, average/minimum pair margin' README.md \
+  || ! grep -Fq 'Markdown frontier artifacts include a `Triggers` column' README.md \
+  || ! grep -Fq 'pair arm, margin, wall ratio, run id, verdict, and trigger reasons' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'with pair arm, verdict, and trigger reasons from the frontier step' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'pair-candidate-frontier.py --fail-on-unmeasured' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'audit-headroom-rejections.py' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'audit-headroom-rejections.py' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'unrecorded headroom rejection(s)' benchmark/auto-resolve/scripts/audit-headroom-rejections.py \
+  || ! grep -Fq 'unsupported registry rejection(s)' benchmark/auto-resolve/scripts/audit-headroom-rejections.py \
+  || ! grep -Fq 'solo_claude={solo_score}' benchmark/auto-resolve/scripts/audit-headroom-rejections.py \
+  || ! grep -Fq 'expected_solo_claude={expected_solo_score}' benchmark/auto-resolve/scripts/audit-headroom-rejections.py \
+  || ! grep -Fq 'unsupported_registry_rejections' benchmark/auto-resolve/scripts/test-audit-headroom-rejections.sh \
+  || ! grep -Fq 'expected_solo_claude=98' benchmark/auto-resolve/scripts/test-audit-headroom-rejections.sh \
+  || ! grep -Fq 'active registry entries whose reason cites a run id or' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'active rejected-registry reason is backed' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'invalid headroom rejections' bin/devlyn.js \
+  || ! grep -Fq 'MALFORMED_JSON' benchmark/auto-resolve/scripts/audit-headroom-rejections.py \
+  || ! grep -Fq 'MALFORMED_ROWS' benchmark/auto-resolve/scripts/audit-headroom-rejections.py \
+  || ! grep -Fq 'F33-cli-new-candidate' benchmark/auto-resolve/scripts/test-audit-headroom-rejections.sh \
+  || ! grep -Fq 'bad-json-headroom <unknown>' benchmark/auto-resolve/scripts/test-audit-headroom-rejections.sh \
+  || ! grep -Fq 'malformed-headroom <unknown>' benchmark/auto-resolve/scripts/test-audit-headroom-rejections.sh \
+  || ! grep -Fq 'candidate_unmeasured' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'pair_evidence_passed' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'print_summary(report)' benchmark/auto-resolve/scripts/pair-candidate-frontier.py \
+  || ! grep -Fq 'fail_on_unmeasured' benchmark/auto-resolve/scripts/pair-candidate-frontier.py \
+  || ! grep -Fq 'unmeasured candidate fixture(s)' benchmark/auto-resolve/scripts/test-pair-candidate-frontier.sh \
+  || ! grep -Fq 'pure JSON stdout must not include final text verdict' benchmark/auto-resolve/scripts/test-pair-candidate-frontier.sh \
+  || ! grep -Fq 'benchmark frontier pure JSON stdout must not include final text verdict' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq -- '--fail-on-unmeasured' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq -- '--fail-on-unmeasured' bin/devlyn.js \
+  || ! grep -Fq -- '--fail-on-unmeasured' benchmark/auto-resolve/README.md \
+  || ! grep -Fq -- '--fail-on-unmeasured' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'bare={bare} solo_claude={solo} pair={pair} arm={arm} margin={margin}' benchmark/auto-resolve/scripts/pair-candidate-frontier.py \
+  || ! grep -Fq 'verdict=pair_evidence_passed' benchmark/auto-resolve/scripts/pair-candidate-frontier.py \
+  || ! grep -Fq 'F16-cli-quote-tax-rules: bare=50 solo_claude=75 pair=96 arm=l2_risk_probes margin=+21' benchmark/auto-resolve/scripts/test-pair-candidate-frontier.sh \
+  || ! grep -Fq 'verdict=pair_evidence_passed' benchmark/auto-resolve/scripts/test-pair-candidate-frontier.sh \
+  || ! grep -Fq '[audit] frontier' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'fixtures=21 rejected=17 candidates=4 pair_evidence=4 unmeasured=0 verdict=PASS' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'F16-cli-quote-tax-rules: bare=50 solo_claude=75 pair=96 arm=l2_risk_probes margin=+21' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'frontier.stdout' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'headroom-rejections.stdout' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq '"frontier_stdout": "frontier.stdout"' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq '"headroom_rejections_stdout": "headroom-rejections.stdout"' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'def load_headroom_audit_summary' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'def check_headroom_audit_report' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'def print_headroom_rejections_summary' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'headroom_rejections={status} verdict={verdict}' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'headroom_report_status == 0' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq '"report_check_exit_code": headroom_report_status' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'headroom audit unsupported registry rejection count missing or malformed' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'headroom_rejections=PASS verdict=PASS unrecorded=0 unsupported=0' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'headroom_rejections=PASS verdict=PASS unrecorded=0 unsupported=0' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'pair trigger reason rows {len(rows)} do not match summary count {count}' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'pair trigger reasons missing for fixture(s)' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'missing-trigger-reasons' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'malformed-trigger-reason-rows' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'mixed-unknown-trigger-reason-rows' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'pair_trigger_reasons=PASS canonical=2 historical_alias=1 exposed=2 total=2 summary=2 rows_match=true' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'pair_trigger_historical_aliases=F21-cli-scheduler-priority=risk_profile.high_risk' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq '"unsupported_registry_rejection_count": (' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'len(unsupported) if isinstance(unsupported, list) else None' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'headroom-missing-unsupported' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'unsupported_registry_rejection_count' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'unsupported_registry_rejection_count' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq '`checks.headroom_rejections` records child verdict plus unrecorded/unsupported counts' README.md \
+  || ! grep -Fq '`checks.headroom_rejections` records the child verdict plus unrecorded and' benchmark/auto-resolve/README.md \
+  || ! grep -Fq '`checks.headroom_rejections` records child verdict plus unrecorded/unsupported counts' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'def check_frontier_stdout' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq '"final_verdict_rows": final_verdict_rows' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'frontier stdout final verdict row count is not exactly 1' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'frontier stdout missing score row' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'wall={wall} run={run_id}' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'frontier stdout aggregate score row count is not exactly 1' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'frontier stdout summary score row count is not exactly 1' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'frontier stdout aggregate fields malformed' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'frontier stdout score row count' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'summary_count = stdout.splitlines().count(required_summary)' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'aggregate_count = stdout.splitlines().count(required_aggregate)' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'frontier stdout summary counts malformed' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'count_keys = {' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'pair_margin_avg={avg} pair_margin_min={min_margin}' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'def format_decimal_margin' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'def format_wall_ratio' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'frontier stdout check missing summary fields' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'missing-frontier-score-row' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'missing-frontier-aggregate-row' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'duplicate-frontier-summary-row' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'duplicate-frontier-aggregate-row' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'pair_margin_avg=+27.00 pair_margin_min=+21 wall_avg=1.38x wall_max=1.47x' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'partial-frontier-score-row' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'extra-frontier-score-row' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'malformed-frontier-stdout-summary' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'malformed-frontier-stdout-counts' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'malformed-frontier-stdout-aggregate' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'audit["checks"]["frontier_stdout"]["status"] == "PASS"' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'audit["checks"]["frontier_stdout"]["summary_rows"] == 1' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'report["checks"]["frontier_stdout"]["aggregate_rows"] == 1' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'audit["checks"]["frontier_stdout"]["final_verdict_rows"] == 1' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'missing-final-frontier-verdict' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'duplicate-final-frontier-verdict' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'audit["checks"]["frontier_stdout"]["expected_rows"] == len(audit["pair_evidence_rows"])' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'audit["checks"]["frontier_stdout"]["trigger_rows"] == len(audit["pair_evidence_rows"])' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'audit["checks"]["frontier_stdout"]["hypothesis_trigger_rows"] == len(audit["pair_evidence_rows"])' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'audit["checks"]["frontier_stdout"]["trigger_rows_match_count"] is True' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'audit["checks"]["frontier_stdout"]["hypothesis_trigger_rows_match_count"] is True' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'report["checks"]["frontier_stdout"]["trigger_rows"] == 2' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'report["checks"]["frontier_stdout"]["hypothesis_trigger_rows"] == 2' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'report["checks"]["frontier_stdout"]["trigger_rows_match_count"] is True' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'report["checks"]["frontier_stdout"]["hypothesis_trigger_rows_match_count"] is True' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'report["checks"]["frontier_stdout"]["rows_match_count"] is True' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'audit["artifacts"]["frontier_stdout"] == "frontier.stdout"' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'report["artifacts"] == {' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'frontier.stdout' benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh \
+  || ! grep -Fq 'frontier.stdout' benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh \
+  || ! grep -Fq 'frontier summary and an artifact map (`artifacts`)' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'artifact map' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'child stdout/stderr logs' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq '| Fixture | Status | Verdict | Evidence | Pair arm | Triggers | Hypothesis trigger | Bare | Solo_claude | Pair | Margin | Wall ratio | Rejected reason |' benchmark/auto-resolve/scripts/test-pair-candidate-frontier.sh \
+  || ! grep -Fq 'benchmark frontier` also prints a stdout score summary for existing complete pair' README.md \
+  || ! grep -Fq 'plus row-level verdicts' README.md \
+  || ! grep -Fq 'including pair arm, trigger reasons, average/minimum pair margin' README.md \
+  || ! grep -Fq 'Markdown frontier artifacts include a `Triggers` column' README.md \
+  || ! grep -Fq 'Full-pipeline pair gate artifacts record `require_hypothesis_trigger` in JSON' README.md \
+  || ! grep -Fq 'Full-pipeline pair gate artifacts record `require_hypothesis_trigger` in JSON' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'Full-pipeline pair gate artifacts record `require_hypothesis_trigger` in JSON' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'includes a Markdown `Hypothesis trigger` column' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'includes a Markdown `Hypothesis trigger` column' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'pair arm, margin, wall ratio, run id, verdict, and trigger reasons' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'overall verdict plus row-level verdict, pair-arm, and trigger-reason columns' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'with pair arm, verdict, and trigger reasons from the frontier step' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'overall verdict plus row-level verdict, pair-arm, and trigger-reason columns' benchmark/auto-resolve/run-real-benchmark.md \
+  || ! grep -Fq 'complete pair evidence rows' benchmark/auto-resolve/README.md \
+  || ! grep -Fq 'complete pair evidence rows' benchmark/auto-resolve/run-real-benchmark.md; then
+  offenders="${offenders}"$'\n'"benchmark docs and CLI must expose the pair-candidate frontier report before new provider spend"
+fi
+if ! grep -Fq '"benchmark/auto-resolve/scripts/**"' package.json \
+  || ! grep -Fq 'benchmark/auto-resolve/results/20260511-f21-current-riskprobes-v1/full-pipeline-pair-gate.json' package.json \
+  || ! grep -Fq 'benchmark/auto-resolve/results/20260512-f2-medium-headroom/headroom-gate.json' package.json \
+  || ! grep -Fq 'benchmark/auto-resolve/results/20260512-f31-seat-rebalance-headroom/headroom-gate.json' package.json \
+  || ! grep -Fq 'benchmark/auto-resolve/results/20260512-f32-subscription-renewal-headroom/headroom-gate.json' package.json; then
+  offenders="${offenders}"$'\n'"package.json must include benchmark scripts, runner regression tests, current pair evidence, and rejected-headroom evidence in npm packages"
+fi
+if ! grep -Fq '"max_observed_pair_solo_wall_ratio": 2.2506234413965087' benchmark/auto-resolve/results/20260510-f16-f23-f25-combined-proof/full-pipeline-pair-gate.json \
+  || ! grep -Fq '"max_observed_pair_solo_wall_ratio": 1.4728476821192054' benchmark/auto-resolve/results/20260511-f21-current-riskprobes-v1/full-pipeline-pair-gate.json \
+  || ! grep -Fq 'pair_trigger eligible with a canonical reason' benchmark/auto-resolve/results/20260510-f16-f23-f25-combined-proof/full-pipeline-pair-gate.json \
+  || ! grep -Fq 'pair_trigger eligible with a canonical reason' benchmark/auto-resolve/results/20260511-f21-current-riskprobes-v1/full-pipeline-pair-gate.json \
+  || ! grep -Fq 'pair_trigger eligible with canonical reason' benchmark/auto-resolve/results/20260510-f16-f23-f25-combined-proof/full-pipeline-pair-gate.md \
+  || ! grep -Fq 'pair_trigger eligible with canonical reason' benchmark/auto-resolve/results/20260511-f21-current-riskprobes-v1/full-pipeline-pair-gate.md \
+  || ! grep -Fq 'Allowed pair/solo wall ratio: 3.00x' benchmark/auto-resolve/results/20260510-f16-f23-f25-combined-proof/full-pipeline-pair-gate.md \
+  || ! grep -Fq 'Maximum observed pair/solo wall ratio: 2.25x' benchmark/auto-resolve/results/20260510-f16-f23-f25-combined-proof/full-pipeline-pair-gate.md \
+  || ! grep -Fq 'Allowed pair/solo wall ratio: 3.00x' benchmark/auto-resolve/results/20260511-f21-current-riskprobes-v1/full-pipeline-pair-gate.md \
+  || ! grep -Fq 'Maximum observed pair/solo wall ratio: 1.47x' benchmark/auto-resolve/results/20260511-f21-current-riskprobes-v1/full-pipeline-pair-gate.md; then
+  offenders="${offenders}"$'\n'"packaged pair evidence artifacts must use the current observed-vs-allowed wall-ratio schema and canonical trigger rule wording"
+fi
+if make_temp_dir package_results /tmp/devlyn-lint-package-results.XXXXXX \
+  && make_temp_dir package_audit /tmp/devlyn-lint-package-audit.XXXXXX \
+  && make_temp_file package_audit_stdout /tmp/devlyn-lint-package-audit.XXXXXX.out; then
+  cp -R benchmark/auto-resolve/results/20260510-f16-f23-f25-combined-proof "$package_results/"
+  cp -R benchmark/auto-resolve/results/20260511-f21-current-riskprobes-v1 "$package_results/"
+  for rejected_run in \
+    20260507-f10-f11-tier1-full-pipeline \
+    20260508-f22-exact-error-headroom \
+    20260508-f26-headroom \
+    20260511-f3-http-error-headroom \
+    20260511-f12-webhook-headroom \
+    20260511-f15-concurrency-headroom \
+    20260512-f2-medium-headroom \
+    20260512-f4-web-headroom \
+    20260512-f5-fixloop-headroom \
+    20260512-f6-checksum-headroom \
+    20260512-f7-scope-headroom \
+    20260512-f9-e2e-headroom \
+    20260512-f31-seat-rebalance-headroom \
+    20260512-f32-subscription-renewal-headroom; do
+    cp -R "benchmark/auto-resolve/results/$rejected_run" "$package_results/"
+  done
+  if ! python3 benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+      --results-root "$package_results" \
+      --out-dir "$package_audit" >"$package_audit_stdout" 2>&1; then
+    offenders="${offenders}"$'\n'"packaged pair evidence subset must pass benchmark audit without relying on unshipped local results"
+  elif ! grep -Fq 'headroom_rejections=PASS verdict=PASS unrecorded=0 unsupported=0' "$package_audit_stdout" \
+    || ! grep -Fq 'pair_evidence_quality=PASS min_pair_margin_actual=+21 min_pair_margin_required=+5 max_wall_actual=2.25x max_wall_allowed=3.00x' "$package_audit_stdout" \
+    || ! grep -Fq 'pair_trigger_reasons=PASS canonical=4 historical_alias=0 exposed=4 total=4 summary=4 rows_match=true' "$package_audit_stdout" \
+    || ! grep -Fq 'pair_evidence_hypothesis_triggers=PASS matched=4 documented=4 total=4' "$package_audit_stdout" \
+    || grep -Fq 'pair_trigger_historical_aliases=' "$package_audit_stdout" \
+    || grep -Fq 'pair_evidence_hypothesis_trigger_gaps=' "$package_audit_stdout"; then
+    offenders="${offenders}"$'\n'"packaged pair evidence subset audit stdout must expose headroom, pair-quality, and trigger-reason handoff rows"
+  elif ! python3 - "$package_audit/audit.json" <<'PY'
+import json
+import sys
+
+report = json.load(open(sys.argv[1], encoding="utf8"))
+rows = report.get("pair_evidence_rows")
+frontier_summary = report.get("frontier_summary", {})
+frontier_report = report.get("checks", {}).get("frontier_report", {})
+frontier_stdout = report.get("checks", {}).get("frontier_stdout", {})
+min_pair_evidence = report.get("checks", {}).get("min_pair_evidence", {})
+pair_evidence_quality = report.get("checks", {}).get("pair_evidence_quality", {})
+pair_trigger_reasons = report.get("checks", {}).get("pair_trigger_reasons", {})
+headroom_rejections = report.get("checks", {}).get("headroom_rejections", {})
+artifacts = report.get("artifacts", {})
+assert report.get("verdict") == "PASS"
+assert frontier_summary.get("verdict") == "PASS"
+assert frontier_summary.get("pair_evidence_count") == 4
+assert frontier_summary.get("unmeasured_count") == 0
+assert frontier_report.get("status") == "PASS"
+assert frontier_report.get("verdict") == "PASS"
+assert frontier_report.get("unmeasured_count") == 0
+assert frontier_stdout.get("status") == "PASS"
+assert frontier_stdout.get("report", "").endswith("frontier.stdout")
+assert frontier_stdout.get("summary_rows") == 1
+assert frontier_stdout.get("aggregate_rows") == 1
+assert frontier_stdout.get("final_verdict_rows") == 1
+assert frontier_stdout.get("expected_rows") == len(rows) == 4
+assert frontier_stdout.get("stdout_rows") == len(rows) == 4
+assert frontier_stdout.get("trigger_rows") == len(rows) == 4
+assert frontier_stdout.get("hypothesis_trigger_rows") == len(rows) == 4
+assert frontier_stdout.get("rows_match_count") is True
+assert frontier_stdout.get("trigger_rows_match_count") is True
+assert frontier_stdout.get("hypothesis_trigger_rows_match_count") is True
+assert headroom_rejections.get("status") == "PASS"
+assert headroom_rejections.get("report_check_exit_code") == 0
+assert headroom_rejections.get("verdict") == "PASS"
+assert headroom_rejections.get("unrecorded_failure_count") == 0
+assert headroom_rejections.get("unsupported_registry_rejection_count") == 0
+assert min_pair_evidence.get("rows_match_count") is True
+assert min_pair_evidence.get("actual_rows") == len(rows) == 4
+assert pair_evidence_quality.get("status") == "PASS"
+assert pair_evidence_quality.get("min_pair_margin_actual") == frontier_summary.get("pair_margin_min")
+assert pair_evidence_quality.get("max_pair_solo_wall_ratio_actual") == frontier_summary.get("pair_solo_wall_ratio_max")
+assert pair_trigger_reasons.get("status") == "PASS"
+assert pair_trigger_reasons.get("summary_pair_evidence_count") == 4
+assert pair_trigger_reasons.get("canonical_rows") == 4
+assert pair_trigger_reasons.get("historical_alias_rows") == 0
+assert pair_trigger_reasons.get("historical_alias_details") == []
+assert pair_trigger_reasons.get("exposed_rows") == 4
+assert pair_trigger_reasons.get("total_rows") == 4
+assert pair_trigger_reasons.get("rows_match_count") is True
+pair_hypothesis_triggers = report.get("checks", {}).get("pair_evidence_hypothesis_triggers", {})
+assert pair_hypothesis_triggers.get("status") == "PASS"
+assert pair_hypothesis_triggers.get("matched_rows") == 4
+assert pair_hypothesis_triggers.get("documented_rows") == 4
+assert pair_hypothesis_triggers.get("gap_details") == []
+assert artifacts.get("frontier_json") == "frontier.json"
+assert artifacts.get("frontier_stdout") == "frontier.stdout"
+assert artifacts.get("headroom_audit_json") == "headroom-audit.json"
+assert artifacts.get("headroom_rejections_stdout") == "headroom-rejections.stdout"
+for row in rows:
+    assert row.get("verdict") == "pair_evidence_passed"
+    assert row.get("pair_arm") == "l2_risk_probes"
+    assert row.get("pair_mode") is True
+    assert row.get("pair_trigger_eligible") is True
+    assert isinstance(row.get("pair_trigger_reasons"), list)
+    assert row.get("pair_trigger_reasons")
+    assert row.get("pair_trigger_has_canonical_reason") is True
+    assert row.get("pair_trigger_has_hypothesis_reason") is True
+PY
+  then
+    offenders="${offenders}"$'\n'"packaged pair evidence subset audit.json must expose frontier_report, frontier_stdout, artifact map, and 4 trigger-backed verdict-bearing pair rows with trigger reasons"
+  fi
+  rm -rf "$package_results" "$package_audit" "$package_audit_stdout"
+else
+  offenders="${offenders}"$'\n'"packaged pair evidence subset could not allocate temporary audit workspace"
+fi
+if make_temp_file pack_json /tmp/devlyn-lint-pack.XXXXXX.json \
+  && npm pack --dry-run --json > "$pack_json" 2>/dev/null; then
+  if ! node - "$pack_json" <<'NODE'
+const fs = require("fs");
+const path = require("path");
+const packPath = process.argv[2];
+const pack = JSON.parse(fs.readFileSync(packPath, "utf8"))[0];
+const files = new Set(pack.files.map((file) => file.path));
+function listFiles(dir) {
+  return fs.readdirSync(dir, { withFileTypes: true }).flatMap((entry) => {
+    const full = path.join(dir, entry.name);
+    if (entry.isDirectory()) return listFiles(full);
+    if (!entry.isFile()) return [];
+    return [full.split(path.sep).join("/")];
+  });
+}
+const shadowRequired = fs
+  .readdirSync("benchmark/auto-resolve/shadow-fixtures", { withFileTypes: true })
+  .filter((entry) => entry.isDirectory() && /^S/.test(entry.name))
+  .flatMap((entry) => listFiles(path.join("benchmark/auto-resolve/shadow-fixtures", entry.name)));
+const required = [
+  "benchmark/auto-resolve/BENCHMARK-RESULTS.md",
+  "benchmark/auto-resolve/run-real-benchmark.md",
+  "benchmark/auto-resolve/results/20260510-f16-f23-f25-combined-proof/headroom-gate.md",
+  "benchmark/auto-resolve/results/20260510-f16-f23-f25-combined-proof/headroom-gate.json",
+  "benchmark/auto-resolve/results/20260510-f16-f23-f25-combined-proof/full-pipeline-pair-gate.md",
+  "benchmark/auto-resolve/results/20260510-f16-f23-f25-combined-proof/full-pipeline-pair-gate.json",
+  "benchmark/auto-resolve/results/20260511-f21-current-riskprobes-v1/headroom-gate.md",
+  "benchmark/auto-resolve/results/20260511-f21-current-riskprobes-v1/headroom-gate.json",
+  "benchmark/auto-resolve/results/20260511-f21-current-riskprobes-v1/full-pipeline-pair-gate.md",
+  "benchmark/auto-resolve/results/20260511-f21-current-riskprobes-v1/full-pipeline-pair-gate.json",
+  "benchmark/auto-resolve/results/20260512-f2-medium-headroom/headroom-gate.json",
+  "benchmark/auto-resolve/results/20260512-f31-seat-rebalance-headroom/headroom-gate.json",
+  "benchmark/auto-resolve/results/20260512-f32-subscription-renewal-headroom/headroom-gate.json",
+  "benchmark/auto-resolve/fixtures/F31-cli-seat-rebalance/NOTES.md",
+	  "benchmark/auto-resolve/fixtures/F31-cli-seat-rebalance/spec.md",
+	  "benchmark/auto-resolve/fixtures/F31-cli-seat-rebalance/expected.json",
+	  "benchmark/auto-resolve/fixtures/F31-cli-seat-rebalance/metadata.json",
+	  "benchmark/auto-resolve/fixtures/F31-cli-seat-rebalance/verifiers/priority-transfer-rollback.js",
+	  "benchmark/auto-resolve/fixtures/F31-cli-seat-rebalance/verifiers/duplicate-event-error.js",
+	  "benchmark/auto-resolve/fixtures/F32-cli-subscription-renewal/NOTES.md",
+	  "benchmark/auto-resolve/fixtures/F32-cli-subscription-renewal/spec.md",
+	  "benchmark/auto-resolve/fixtures/F32-cli-subscription-renewal/expected.json",
+	  "benchmark/auto-resolve/fixtures/F32-cli-subscription-renewal/metadata.json",
+	  "benchmark/auto-resolve/fixtures/F32-cli-subscription-renewal/verifiers/priority-credit-rollback.js",
+	  "benchmark/auto-resolve/fixtures/F32-cli-subscription-renewal/verifiers/duplicate-renewal-error.js",
+	  "benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh",
+  "benchmark/auto-resolve/scripts/pair-rejected-fixtures.sh",
+  "benchmark/auto-resolve/scripts/pair_evidence_contract.py",
+  "benchmark/auto-resolve/scripts/pair-candidate-frontier.py",
+  "benchmark/auto-resolve/scripts/test-pair-candidate-frontier.sh",
+  "benchmark/auto-resolve/scripts/audit-pair-evidence.py",
+  "benchmark/auto-resolve/scripts/solo-headroom-hypothesis.py",
+  "benchmark/auto-resolve/scripts/solo-ceiling-avoidance.py",
+  "benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh",
+  "benchmark/auto-resolve/scripts/test-build-pair-eligible-manifest.sh",
+  "benchmark/auto-resolve/scripts/test-check-f9-artifacts.sh",
+  "benchmark/auto-resolve/scripts/test-lint-fixtures.sh",
+  "benchmark/auto-resolve/scripts/test-run-full-pipeline-pair-candidate.sh",
+	  "benchmark/auto-resolve/scripts/test-run-headroom-candidate.sh",
+  "benchmark/auto-resolve/scripts/test-run-swebench-solver-batch.sh",
+	  "benchmark/auto-resolve/scripts/test-ship-gate.sh",
+	  "benchmark/auto-resolve/scripts/test-iter-0033c-l1-summary.sh",
+  "benchmark/auto-resolve/scripts/test-iter-0033c-compare.sh",
+  "benchmark/auto-resolve/fixtures/retired/F27-cli-subscription-proration/RETIRED.md",
+  "benchmark/auto-resolve/fixtures/retired/F27-cli-subscription-proration/spec.md",
+  "benchmark/auto-resolve/fixtures/retired/F28-cli-return-authorization/RETIRED.md",
+  "benchmark/auto-resolve/fixtures/retired/F28-cli-return-authorization/spec.md",
+  "benchmark/auto-resolve/fixtures/retired/F30-cli-credit-hold-settlement/RETIRED.md",
+  "benchmark/auto-resolve/fixtures/retired/F30-cli-credit-hold-settlement/spec.md",
+  "benchmark/auto-resolve/fixtures/retired/F9-e2e-ideate-to-preflight/RETIRED.md",
+  "scripts/lint-fixtures.sh",
+  "scripts/lint-shadow-fixtures.sh",
+  ...shadowRequired,
+];
+const missing = required.filter((file) => !files.has(file));
+if (missing.length > 0) {
+  console.error(missing.join("\n"));
+  process.exit(1);
+}
+const forbidden = pack.files
+  .map((file) => file.path)
+  .filter((file) => file.includes("__pycache__") || file.endsWith(".pyc"));
+if (forbidden.length > 0) {
+  console.error(forbidden.join("\n"));
+  process.exit(2);
+}
+NODE
+  then
+    offenders="${offenders}"$'\n'"npm pack dry-run must include benchmark runner/gate regression tests, all shadow fixture files, retired fixture replay docs, and exclude pycache artifacts"
+  fi
+else
+  offenders="${offenders}"$'\n'"npm pack dry-run failed while checking benchmark runner/gate regression tests"
+fi
+[ -n "${pack_json:-}" ] && rm -f "$pack_json"
+non_executable_shell_scripts=$(find benchmark/auto-resolve/scripts -maxdepth 1 -name '*.sh' -type f ! -perm -111 -print | sort)
+if [ -n "$non_executable_shell_scripts" ]; then
+  while IFS= read -r f; do
+    offenders="${offenders}"$'\n'"benchmark shell script must be executable: $f"
+  done <<< "$non_executable_shell_scripts"
+fi
+for gate_test in \
+  test-benchmark-arg-parsing.sh \
+  test-pair-candidate-frontier.sh \
+  test-audit-pair-evidence.sh \
+  test-audit-headroom-rejections.sh \
+  test-build-pair-eligible-manifest.sh \
+  test-ship-gate.sh \
+  test-headroom-gate.sh \
+  test-run-headroom-candidate.sh \
+  test-check-f9-artifacts.sh \
+  test-lint-fixtures.sh \
+	  test-run-full-pipeline-pair-candidate.sh \
+	  test-full-pipeline-pair-gate.sh \
+	  test-iter-0033c-l1-summary.sh \
+	  test-iter-0033c-compare.sh \
+  test-run-swebench-solver-batch.sh \
+  test-swebench-frozen-case.sh \
+  test-frozen-verify-gate.sh
+do
+  if ! grep -Fq "$gate_test" benchmark/auto-resolve/README.md; then
+    offenders="${offenders}"$'\n'"benchmark/auto-resolve/README.md: gate-change instructions must list $gate_test"
+  fi
+done
+if ! bash benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh >/dev/null 2>&1; then
+  offenders="${offenders}"$'\n'"benchmark/auto-resolve/scripts/test-benchmark-arg-parsing.sh: failed"
+fi
+if ! bash benchmark/auto-resolve/scripts/test-pair-candidate-frontier.sh >/dev/null 2>&1; then
+  offenders="${offenders}"$'\n'"benchmark/auto-resolve/scripts/test-pair-candidate-frontier.sh: failed"
+fi
+if ! bash benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh >/dev/null 2>&1; then
+  offenders="${offenders}"$'\n'"benchmark/auto-resolve/scripts/test-audit-pair-evidence.sh: failed"
+fi
+if ! bash benchmark/auto-resolve/scripts/test-audit-headroom-rejections.sh >/dev/null 2>&1; then
+  offenders="${offenders}"$'\n'"benchmark/auto-resolve/scripts/test-audit-headroom-rejections.sh: failed"
+fi
+if ! bash benchmark/auto-resolve/scripts/test-build-pair-eligible-manifest.sh >/dev/null 2>&1; then
+  offenders="${offenders}"$'\n'"benchmark/auto-resolve/scripts/test-build-pair-eligible-manifest.sh: failed"
+fi
+if ! bash benchmark/auto-resolve/scripts/test-headroom-gate.sh >/dev/null 2>&1; then
+  offenders="${offenders}"$'\n'"benchmark/auto-resolve/scripts/test-headroom-gate.sh: failed"
+fi
+if ! bash benchmark/auto-resolve/scripts/test-run-headroom-candidate.sh >/dev/null 2>&1; then
+  offenders="${offenders}"$'\n'"benchmark/auto-resolve/scripts/test-run-headroom-candidate.sh: failed"
+fi
+if ! bash benchmark/auto-resolve/scripts/test-check-f9-artifacts.sh >/dev/null 2>&1; then
+  offenders="${offenders}"$'\n'"benchmark/auto-resolve/scripts/test-check-f9-artifacts.sh: failed"
+fi
+if ! bash benchmark/auto-resolve/scripts/test-run-full-pipeline-pair-candidate.sh >/dev/null 2>&1; then
+  offenders="${offenders}"$'\n'"benchmark/auto-resolve/scripts/test-run-full-pipeline-pair-candidate.sh: failed"
+fi
+if ! bash benchmark/auto-resolve/scripts/test-ship-gate.sh >/dev/null 2>&1; then
+  offenders="${offenders}"$'\n'"benchmark/auto-resolve/scripts/test-ship-gate.sh: failed"
+fi
+if ! bash benchmark/auto-resolve/scripts/test-full-pipeline-pair-gate.sh >/dev/null 2>&1; then
+  offenders="${offenders}"$'\n'"benchmark/auto-resolve/scripts/test-full-pipeline-pair-gate.sh: failed"
+fi
+if ! bash benchmark/auto-resolve/scripts/test-iter-0033c-l1-summary.sh >/dev/null 2>&1; then
+  offenders="${offenders}"$'\n'"benchmark/auto-resolve/scripts/test-iter-0033c-l1-summary.sh: failed"
+fi
+if ! bash benchmark/auto-resolve/scripts/test-iter-0033c-compare.sh >/dev/null 2>&1; then
+  offenders="${offenders}"$'\n'"benchmark/auto-resolve/scripts/test-iter-0033c-compare.sh: failed"
+fi
+if ! bash benchmark/auto-resolve/scripts/test-run-swebench-solver-batch.sh >/dev/null 2>&1; then
+  offenders="${offenders}"$'\n'"benchmark/auto-resolve/scripts/test-run-swebench-solver-batch.sh: failed"
+fi
+if ! bash benchmark/auto-resolve/scripts/test-swebench-frozen-case.sh >/dev/null 2>&1; then
+  offenders="${offenders}"$'\n'"benchmark/auto-resolve/scripts/test-swebench-frozen-case.sh: failed"
+fi
+if ! bash benchmark/auto-resolve/scripts/test-frozen-verify-gate.sh >/dev/null 2>&1; then
+  offenders="${offenders}"$'\n'"benchmark/auto-resolve/scripts/test-frozen-verify-gate.sh: failed"
+fi
+if ! bash scripts/lint-fixtures.sh >/dev/null 2>&1; then
+  offenders="${offenders}"$'\n'"scripts/lint-fixtures.sh: failed"
+fi
+if ! bash scripts/lint-shadow-fixtures.sh >/dev/null 2>&1; then
+  offenders="${offenders}"$'\n'"scripts/lint-shadow-fixtures.sh: failed"
+fi
+if ! bash benchmark/auto-resolve/scripts/test-lint-fixtures.sh >/dev/null 2>&1; then
+  offenders="${offenders}"$'\n'"benchmark/auto-resolve/scripts/test-lint-fixtures.sh: failed"
+fi
+if ! python3 -m py_compile benchmark/auto-resolve/scripts/solo-headroom-hypothesis.py >/dev/null 2>&1; then
+  offenders="${offenders}"$'\n'"benchmark/auto-resolve/scripts/solo-headroom-hypothesis.py: py_compile failed"
+fi
+if ! grep -Fq 'high-risk fixture must include a resolve risk-trigger term' scripts/lint-fixtures.sh \
+  || ! grep -Fq 'SOLO_HEADROOM_CHECK="$REPO_ROOT/benchmark/auto-resolve/scripts/solo-headroom-hypothesis.py"' scripts/lint-fixtures.sh \
+  || ! grep -Fq 'pair_evidence_passed fixture spec.md must document an actionable solo-headroom hypothesis with solo_claude miss and observable command from expected.json' scripts/lint-fixtures.sh \
+  || ! grep -Fq -- '--expected-json "$d/expected.json" "$d/spec.md"' scripts/lint-fixtures.sh \
+  || ! grep -Fq 'str(fixture_dir / "expected.json")' benchmark/auto-resolve/scripts/audit-pair-evidence.py \
+  || ! grep -Fq 'hypothesis command must match expected.json verification command' benchmark/auto-resolve/scripts/test-lint-fixtures.sh \
+  || ! grep -Fq 'pair-evidence-hypothesis-fail.out' benchmark/auto-resolve/scripts/test-lint-fixtures.sh \
+  || ! grep -Fq 'pair-evidence-hypothesis-pass.out' benchmark/auto-resolve/scripts/test-lint-fixtures.sh \
+  || ! grep -Fq 'DEVLYN_FIXTURES_DIR' scripts/lint-fixtures.sh \
+  || ! grep -Fq 'DEVLYN_FIXTURE_GLOB' scripts/lint-fixtures.sh \
+  || ! grep -Fq 'DEVLYN_FIXTURE_GLOB="S*"' scripts/lint-shadow-fixtures.sh \
+  || ! grep -Fq 'benchmark/auto-resolve/shadow-fixtures' scripts/lint-shadow-fixtures.sh \
+  || ! grep -Fq 'category") != "high-risk"' scripts/lint-fixtures.sh \
+  || ! grep -Fq 'permissions?' scripts/lint-fixtures.sh \
+  || ! grep -Fq 'idempoten\w*' scripts/lint-fixtures.sh \
+  || ! grep -Fq 'output-shape' scripts/lint-fixtures.sh \
+  || ! grep -Fq 'high-risk trigger validation' benchmark/auto-resolve/scripts/test-lint-fixtures.sh \
+  || ! grep -Fq 'DEVLYN_FIXTURES_DIR="$FIXTURES_DIR" bash "$ROOT/scripts/lint-fixtures.sh"' benchmark/auto-resolve/scripts/test-lint-fixtures.sh \
+  || ! grep -Fq 'DEVLYN_LINT_FIXTURES_NO_JSONSCHEMA=1' benchmark/auto-resolve/scripts/test-lint-fixtures.sh \
+	  || ! grep -Fq 'spec-verify-check --check-expected failed' benchmark/auto-resolve/scripts/test-lint-fixtures.sh \
+	  || ! grep -Fq 'unless sibling spec.md declares all Requirements are pure-design' benchmark/auto-resolve/scripts/test-lint-fixtures.sh \
+	  || ! grep -Fq 'verification_commands must be an array' benchmark/auto-resolve/scripts/test-lint-fixtures.sh \
+	  || ! grep -Fq 'expected.json must be an object' benchmark/auto-resolve/scripts/test-lint-fixtures.sh \
+	  || ! grep -Fq 'hidden oracle missing contract_refs' benchmark/auto-resolve/scripts/test-lint-fixtures.sh \
+	  || ! grep -Fq 'contract_ref not found in spec.md' benchmark/auto-resolve/scripts/test-lint-fixtures.sh \
+	  || ! grep -Fq 'BENCH_FIXTURE_DIR file not found' benchmark/auto-resolve/scripts/test-lint-fixtures.sh \
+	  || ! grep -Fq 'BENCH_FIXTURE_DIR file escapes fixture dir' benchmark/auto-resolve/scripts/test-lint-fixtures.sh \
+	  || ! grep -Fq 'hidden oracle must reference an explicit $BENCH_FIXTURE_DIR/... file' benchmark/auto-resolve/scripts/test-lint-fixtures.sh \
+	  || ! grep -Fq 'hidden oracle must assert stdout_contains includes' benchmark/auto-resolve/scripts/test-lint-fixtures.sh \
+	  || ! grep -Fq 'cd \"$BENCH_FIXTURE_DIR\" && node verifiers/hidden-oracle.js' benchmark/auto-resolve/scripts/test-lint-fixtures.sh \
+	  || ! grep -Fq 'missing-hidden-oracle.js' benchmark/auto-resolve/scripts/test-lint-fixtures.sh \
+	  || ! grep -Fq '../outside-hidden-oracle.js' benchmark/auto-resolve/scripts/test-lint-fixtures.sh \
+	  || ! grep -Fq 'This visible contract is not in the spec.' benchmark/auto-resolve/scripts/test-lint-fixtures.sh \
+	  || ! grep -Fq 'spec-verify-check --check failed' benchmark/auto-resolve/scripts/test-lint-fixtures.sh \
+	  || ! grep -Fq 'frontmatter complexity must be one of' benchmark/auto-resolve/scripts/test-lint-fixtures.sh \
+	  || ! grep -Fq 'resolve spec contract enum' benchmark/auto-resolve/fixtures/SCHEMA.md \
+	  || ! grep -Fq 'metadata.difficulty' benchmark/auto-resolve/fixtures/SCHEMA.md \
+	  || ! grep -Fq "grep -Fq 'Traceback'" benchmark/auto-resolve/scripts/test-lint-fixtures.sh \
+	  || ! grep -Fq 'hidden oracle missing contract_refs' scripts/lint-fixtures.sh \
+	  || ! grep -Fq 'contract_ref not found in spec.md' scripts/lint-fixtures.sh \
+	  || ! grep -Fq 'BENCH_FIXTURE_DIR file not found' scripts/lint-fixtures.sh \
+	  || ! grep -Fq 'BENCH_FIXTURE_DIR file escapes fixture dir' scripts/lint-fixtures.sh \
+	  || ! grep -Fq 'hidden oracle must reference an explicit $BENCH_FIXTURE_DIR/... file' scripts/lint-fixtures.sh \
+	  || ! grep -Fq 'hidden oracle must assert stdout_contains includes' scripts/lint-fixtures.sh \
+	  || ! grep -Fq 'SPEC_VERIFY_CHECK="$REPO_ROOT/config/skills/_shared/spec-verify-check.py"' scripts/lint-fixtures.sh \
+	  || ! grep -Fq 'python3 "$SPEC_VERIFY_CHECK" --check "$d/spec.md"' scripts/lint-fixtures.sh \
+	  || ! grep -Fq 'spec-verify-check --check failed' scripts/lint-fixtures.sh \
+	  || ! grep -Fq 'python3 "$SPEC_VERIFY_CHECK" --check-expected "$d/expected.json"' scripts/lint-fixtures.sh \
+	  || ! grep -Fq 'spec-verify-check --check-expected failed' scripts/lint-fixtures.sh \
+	  || ! grep -Fq 'REJECTED_REGISTRY="${DEVLYN_REJECTED_FIXTURE_REGISTRY:-$REPO_ROOT/benchmark/auto-resolve/scripts/pair-rejected-fixtures.sh}"' scripts/lint-fixtures.sh \
+	  || ! grep -Fq 'declare -F rejected_pair_fixture_reason' scripts/lint-fixtures.sh \
+	  || ! grep -Fq 'rejected fixture registry must define rejected_pair_fixture_reason' benchmark/auto-resolve/scripts/test-lint-fixtures.sh \
+	  || ! grep -Fq 'malformed-rejected.sh' benchmark/auto-resolve/scripts/test-lint-fixtures.sh \
+	  || ! grep -Fq 'NOTES.md records pair-candidate rejection but pair-rejected-fixtures.sh has no rejected reason' scripts/lint-fixtures.sh \
+	  || ! grep -Fq 'test-active-headroom' benchmark/auto-resolve/scripts/test-lint-fixtures.sh \
+	  || ! grep -Fq 'active-calibration-rejected-missing.out' benchmark/auto-resolve/scripts/test-lint-fixtures.sh \
+	  || ! grep -Fq 'Rejected controls should remain replayable' benchmark/auto-resolve/fixtures/SCHEMA.md \
+	  || ! grep -Fq 'pair-rejected-fixtures.sh' benchmark/auto-resolve/fixtures/SCHEMA.md \
+	  || ! grep -Fq 'target.relative_to(fixture_root)' scripts/lint-fixtures.sh \
+	  || ! grep -Fq '$BENCH_FIXTURE_DIR/...' benchmark/auto-resolve/fixtures/SCHEMA.md \
+	  || ! grep -Fq 'must not escape the fixture directory' benchmark/auto-resolve/fixtures/SCHEMA.md \
+	  || ! grep -Fq 'cd "$BENCH_FIXTURE_DIR"` indirection' benchmark/auto-resolve/fixtures/SCHEMA.md \
+	  || ! grep -Fq 'success sentinel' benchmark/auto-resolve/fixtures/SCHEMA.md \
+	  || ! grep -Fq 'def fallback_validate():' scripts/lint-fixtures.sh \
+  || ! grep -Fq 'expected.json must be an object' scripts/lint-fixtures.sh \
+  || ! grep -Fq 'schema_ok=0' scripts/lint-fixtures.sh \
+  || ! grep -Fq 'DEVLYN_LINT_FIXTURES_NO_JSONSCHEMA' scripts/lint-fixtures.sh \
+  || ! grep -Fq 'conditional pair/risk-probe triggers' benchmark/auto-resolve/fixtures/SCHEMA.md; then
+  offenders="${offenders}"$'\n'"fixture lint must require high-risk fixtures to include resolve pair/risk trigger terms"
+fi
+unsafe_json_parser_refs=$(python3 - <<'PY'
+import pathlib
+import re
+
+root = pathlib.Path("benchmark/auto-resolve/scripts")
+patterns = [
+    re.compile(r"json\.load\(open\("),
+    re.compile(r"json\.load\(response\)"),
+    re.compile(r"json\.loads\([^,\n]*(?:read_text\(\)|\bline\b|\bln\b)[^,\n]*\)"),
+]
+for path in sorted(root.glob("*")):
+    if path.name.startswith("test-") or path.suffix not in {".py", ".sh"}:
+        continue
+    text = path.read_text(errors="ignore")
+    for line_no, line in enumerate(text.splitlines(), 1):
+        if any(pattern.search(line) for pattern in patterns):
+            print(f"{path}:{line_no}:{line.strip()}")
+PY
+)
+if [ -n "$unsafe_json_parser_refs" ]; then
+  while IFS= read -r f; do
+    offenders="${offenders}"$'\n'"benchmark script uses unsafe JSON parser without strict constants: $f"
+  done <<< "$unsafe_json_parser_refs"
+fi
+if [ -z "$offenders" ]; then
+  ok "benchmark docs use current bare / solo_claude / pair-arm topology"
+else
+  while IFS= read -r f; do bad "$f"; done <<< "$offenders"
+fi
+
+# ---------------------------------------------------------------------------
+# 9. Engine availability fails closed; stale silent-downgrade wording is forbidden.
 # ---------------------------------------------------------------------------
-section "Check 9: Downgrade string uses 'codex-unavailable'"
-offenders=$(grep -RIln 'codex-ping failed\|codex-ping fail' \
-  config/skills CLAUDE.md README.md bin/ 2>/dev/null \
+section "Check 9: Engine availability fails closed"
+offenders=$(grep -RInE 'codex-ping failed|codex-ping fail|engine downgraded: codex-unavailable|downgrades to Claude-only|silently downgrades|silently downgrade|silently switch to Claude|Codex CLI availability downgrade' \
+  config/skills CLAUDE.md README.md bin/ benchmark/auto-resolve/run-real-benchmark.md 2>/dev/null \
   | grep -v 'roadmap-archival-workspace/' \
   | grep -v 'devlyn:auto-resolve-workspace/' \
   | grep -v 'devlyn:ideate-workspace/' \
   | grep -v 'preflight-workspace/' \
   || true)
 if [ -z "$offenders" ]; then
-  ok "all downgrade strings canonical"
+  ok "engine availability fail-closed wording canonical"
 else
   while IFS= read -r f; do bad "$f"; done <<< "$offenders"
 fi
 
+# ---------------------------------------------------------------------------
+# 9b. Release tag and package version parity.
+#     v2.2.3 was tagged while package.json still said 2.2.2. A mismatched npm
+#     package makes Codex/Claude installed-skill drift hard to diagnose because
+#     users can be on a tag whose package metadata points at a different build.
+# ---------------------------------------------------------------------------
+section "Check 9b: package.json version matches exact release tag"
+exact_tag=$(git describe --tags --exact-match HEAD 2>/dev/null || true)
+pkg_version=$(node -p "require('./package.json').version" 2>/dev/null || true)
+if [ -z "$exact_tag" ]; then
+  ok "HEAD is not an exact tag — package/tag parity not applicable"
+elif [ "$pkg_version" = "${exact_tag#v}" ]; then
+  ok "package.json version matches $exact_tag"
+else
+  bad "package.json version '$pkg_version' does not match HEAD tag '$exact_tag'"
+fi
+
 # ---------------------------------------------------------------------------
 # (Check 7 retired iter-0034 Phase 4 cutover: the 4 findings-producing
 # standalones — evaluate / review / clean / team-review — were deleted; the
@@ -368,29 +3832,31 @@ else
 
   # Content: byte-compare each section block via diff over temp files.
   # awk-into-tmpfile preserves trailing newlines (command substitution strips them).
-  tmp_rp=$(mktemp)
-  tmp_claude=$(mktemp)
-  for name in $expected_sections; do
-    begin="<!-- runtime-principles:section=${name}:begin -->"
-    end="<!-- runtime-principles:section=${name}:end -->"
-    awk -v b="$begin" -v e="$end" '$0==b{f=1;next}$0==e{f=0}f' "$rp_src" > "$tmp_rp"
-    awk -v b="$begin" -v e="$end" '$0==b{f=1;next}$0==e{f=0}f' "$claude_src" > "$tmp_claude"
-    if [ ! -s "$tmp_rp" ]; then
-      bad "${name}: empty/missing block in $rp_src"
-      rp_drift=1
-      continue
-    fi
-    if [ ! -s "$tmp_claude" ]; then
-      bad "${name}: empty/missing block in $claude_src"
-      rp_drift=1
-      continue
-    fi
-    if ! diff -q "$tmp_rp" "$tmp_claude" >/dev/null 2>&1; then
-      bad "${name}: CLAUDE.md and runtime-principles.md content differ"
-      rp_drift=1
-    fi
-  done
-  rm -f "$tmp_rp" "$tmp_claude"
+  if make_temp_file tmp_rp && make_temp_file tmp_claude; then
+    for name in $expected_sections; do
+      begin="<!-- runtime-principles:section=${name}:begin -->"
+      end="<!-- runtime-principles:section=${name}:end -->"
+      awk -v b="$begin" -v e="$end" '$0==b{f=1;next}$0==e{f=0}f' "$rp_src" > "$tmp_rp"
+      awk -v b="$begin" -v e="$end" '$0==b{f=1;next}$0==e{f=0}f' "$claude_src" > "$tmp_claude"
+      if [ ! -s "$tmp_rp" ]; then
+        bad "${name}: empty/missing block in $rp_src"
+        rp_drift=1
+        continue
+      fi
+      if [ ! -s "$tmp_claude" ]; then
+        bad "${name}: empty/missing block in $claude_src"
+        rp_drift=1
+        continue
+      fi
+      if ! diff -q "$tmp_rp" "$tmp_claude" >/dev/null 2>&1; then
+        bad "${name}: CLAUDE.md and runtime-principles.md content differ"
+        rp_drift=1
+      fi
+    done
+    rm -f "$tmp_rp" "$tmp_claude"
+  else
+    rp_drift=1
+  fi
 
   if [ $rp_drift -eq 0 ]; then
     ok "all 4 contract sections in parity (subtractive-first / goal-locked / no-workaround / evidence) — markers, topology, content"
@@ -411,20 +3877,44 @@ if [ ! -x "$idgen" ] && [ ! -f "$idgen" ]; then
 elif [ ! -d "$fixture" ]; then
   bad "Check 13 prerequisite missing: $fixture"
 else
-  tmp1=$(mktemp); tmp2=$(mktemp)
-  if python3 "$idgen" --fixture "$fixture" --generated-at 2026-04-29T18:30:00Z --output "$tmp1" >/dev/null 2>&1 \
-     && python3 "$idgen" --fixture "$fixture" --generated-at 2026-04-29T18:30:00Z --output "$tmp2" >/dev/null 2>&1; then
-    sha1=$(shasum -a 256 "$tmp1" | awk '{print $1}')
-    sha2=$(shasum -a 256 "$tmp2" | awk '{print $1}')
-    if [ "$sha1" = "$sha2" ]; then
-      ok "F2 registry sha256 stable across two idgen runs ($sha1)"
+  if make_temp_file tmp1 && make_temp_file tmp2; then
+    if python3 "$idgen" --fixture "$fixture" --generated-at 2026-04-29T18:30:00Z --output "$tmp1" >/dev/null 2>&1 \
+       && python3 "$idgen" --fixture "$fixture" --generated-at 2026-04-29T18:30:00Z --output "$tmp2" >/dev/null 2>&1; then
+      sha1=$(shasum -a 256 "$tmp1" | awk '{print $1}')
+      sha2=$(shasum -a 256 "$tmp2" | awk '{print $1}')
+      if [ "$sha1" = "$sha2" ]; then
+        ok "F2 registry sha256 stable across two idgen runs ($sha1)"
+      else
+        bad "F2 registry sha256 drift: run1=$sha1 run2=$sha2"
+      fi
     else
-      bad "F2 registry sha256 drift: run1=$sha1 run2=$sha2"
+      bad "idgen invocation failed; cannot verify determinism"
     fi
   else
-    bad "idgen invocation failed; cannot verify determinism"
+    bad "idgen temp files unavailable; cannot verify determinism"
   fi
-  rm -f "$tmp1" "$tmp2"
+  if make_temp_dir tmp_bad && make_temp_file idgen_nan_out /tmp/pair-plan-idgen-nan.XXXXXX.out; then
+    cp -R "$fixture/." "$tmp_bad/"
+    printf '{"verification_commands": NaN}\n' > "$tmp_bad/expected.json"
+    if python3 "$idgen" --fixture "$tmp_bad" --generated-at 2026-04-29T18:30:00Z >"$idgen_nan_out" 2>&1; then
+      bad "pair-plan-idgen.py accepted NaN in expected.json"
+    elif ! grep -Fq 'invalid JSON numeric constant: NaN' "$idgen_nan_out"; then
+      bad "pair-plan-idgen.py NaN failure did not cite invalid JSON numeric constant"
+    fi
+    rm -rf "$tmp_bad" "$idgen_nan_out"
+  fi
+
+  if make_temp_file bad_plan && make_temp_file plan_lint_nan_out /tmp/pair-plan-lint-nan.XXXXXX.out; then
+    printf '{"schema_version": NaN}\n' > "$bad_plan"
+    if python3 benchmark/auto-resolve/scripts/pair-plan-lint.py --plan "$bad_plan" --quiet >"$plan_lint_nan_out" 2>&1; then
+      bad "pair-plan-lint.py accepted NaN in pair-plan.json"
+    elif ! grep -Fq '"code": "plan_invalid_json"' "$plan_lint_nan_out" \
+         || ! grep -Fq 'invalid JSON numeric constant: NaN' "$plan_lint_nan_out"; then
+      bad "pair-plan-lint.py NaN failure did not report plan_invalid_json with invalid numeric constant"
+    fi
+    rm -f "$bad_plan" "$plan_lint_nan_out"
+  fi
+  rm -f "${tmp1:-}" "${tmp2:-}"
 fi
 
 # ---------------------------------------------------------------------------
@@ -467,6 +3957,66 @@ if [ $f9_drift -eq 0 ]; then
   ok "F9 fixture id is canonical (F9-e2e-ideate-to-resolve); no stale refs outside retired/"
 fi
 
+# ---------------------------------------------------------------------------
+# 15. Current pair-evidence fixtures carry their local evidence handoff notes.
+#     The audit artifacts are canonical, but fixture-level notes keep future
+#     fixture edits from losing why a candidate currently counts as solo<pair
+#     evidence.
+# ---------------------------------------------------------------------------
+section "Check 15: Pair evidence fixture notes cite current passing runs"
+if python3 - <<'PY'
+import importlib.util
+import pathlib
+import sys
+
+script = pathlib.Path("benchmark/auto-resolve/scripts/pair-candidate-frontier.py")
+spec = importlib.util.spec_from_file_location("pair_candidate_frontier", script)
+module = importlib.util.module_from_spec(spec)
+assert spec.loader is not None
+spec.loader.exec_module(module)
+
+report = module.build_report(
+    fixtures_root=pathlib.Path("benchmark/auto-resolve/fixtures"),
+    registry=pathlib.Path("benchmark/auto-resolve/scripts/pair-rejected-fixtures.sh"),
+    results_root=pathlib.Path("benchmark/auto-resolve/results"),
+)
+errors = []
+for row in report["rows"]:
+    if row.get("status") != "pair_evidence_passed":
+        continue
+    fixture = row["fixture"]
+    best = module.best_pair_evidence(row.get("passing_pair_evidence", []))
+    if best is None:
+        errors.append(f"{fixture}: missing complete pair evidence")
+        continue
+    notes_path = pathlib.Path("benchmark/auto-resolve/fixtures") / fixture / "NOTES.md"
+    try:
+        notes = notes_path.read_text(encoding="utf8")
+    except OSError:
+        errors.append(f"{fixture}: NOTES.md missing")
+        continue
+    required = [
+        str(best["run_id"]),
+        "pair_evidence_passed",
+        f"bare `{best['bare_score']}`",
+        f"solo_claude `{best['solo_score']}`",
+        f"pair `{best['pair_score']}`",
+        f"margin `{best['pair_margin']:+d}`",
+        f"wall `{best['pair_solo_wall_ratio']:.2f}x`",
+        f"arm `{best['pair_arm']}`",
+    ]
+    missing = [item for item in required if item not in notes]
+    if missing:
+        errors.append(f"{fixture}: NOTES.md missing {', '.join(missing)}")
+if errors:
+    raise SystemExit("\n".join(errors))
+PY
+then
+  ok "current pair-evidence fixture notes cite passing run ids"
+else
+  bad "current pair-evidence fixture NOTES.md files must cite passing run ids and pair_evidence_passed"
+fi
+
 # ---------------------------------------------------------------------------
 # Summary.
 # ---------------------------------------------------------------------------