@event4u/agent-config 14.7.0 → 14.8.0

package/src/config/gate-coverage.yml

@@ -42,16 +42,53 @@
 #
 # COVERAGE — THE HONEST DENOMINATOR (re-measured 2026-08-18)
 # ----------------------------------------------------------
-# 267 `lint_*` / `check_*` / `audit_*` scripts live under `src/scripts/`
-# (the count is computed, not asserted — `check_gate_coverage.count_gate_scripts`).
-# **42 of them emit the machine-readable `scanned: <N>` line under their
-# CI-identical argv, and 41 of those 42 are listed below.** The manifest is
+# 269 `lint_*` / `check_*` / `audit_*` scripts live under `src/scripts/`
+# (the count is computed, not asserted — `check_gate_coverage.count_gate_scripts`,
+# re-run 2026-08-21 and it returns 269).
+# **45 of them emit the machine-readable `scanned: <N>` line under their
+# CI-identical argv, and 42 of those 45 are listed below.** The manifest is
 # therefore complete with respect to its own contract for every gate that can
-# carry a floor, and covers ~15 % of the population.
+# carry a floor, and covers ~16 % of the population.
 #
-# The one emitter deliberately NOT listed is `lint_evidence_artifacts`, for the
-# same reason `check_rule_projection_integrity` is not (see the paragraph
-# below): it is diff-scoped, so it legitimately emits `scanned: 0` on any
+# Both halves were RE-DERIVED on 2026-08-21, not incremented, and the previous
+# reading was off by one in a way worth recording: it said "42 … and 41 of those
+# 42 are listed" while the file carried 42 rows AND named one deliberately
+# unlisted emitter (`lint_evidence_artifacts`, below). Those three facts cannot
+# all hold — 42 listed rows plus 1 unlisted emitter is 43 emitters, not 42. It
+# is the same drift this section documents three times: an entry was added and
+# the prose was adjusted by one instead of recounted.
+#
+# The derivation now, so the next reader can check it rather than trust it:
+# `grep -c '^  - id:'` returns 42 (computable, and computed); every listed row
+# is an emitter, because a row that did not emit could carry no floor and the
+# guard would report it `failing`; THREE emitters are deliberately unlisted —
+# `lint_evidence_artifacts` (below), `adr_cite_check` and
+# `check_new_adr_evidence` (each in the block where its row would have sat);
+# 42 + 3 = 45. The population figure comes from the function, and it stays 269
+# because `check_new_adr_evidence` matches the population prefix whether or not
+# it carries a row here.
+#
+# `adr/evidence_census` emits a `scanned:` line too and is NOT in either number:
+# it carries no `lint|check|audit` prefix, so `count_gate_scripts()` never sees
+# it. Named because the omission looks like a fourth miscount and is not one —
+# an emitter outside the population cannot be an unlisted emitter within it.
+#
+# Corrected in completion review, and it is the third time this paragraph has
+# been off by one — which is the argument for computing it rather than
+# adjusting it. The previous reading said 43 listed and self-invited a `grep`
+# that returns 42, so the check it offered refuted the sentence containing it.
+# The cause each time is the same: a row was added or removed and the prose
+# was incremented instead of recounted.
+#
+# A shape-grep over `src/scripts/` (`reportScanned(` or a literal `scanned: ${`)
+# returns 63 files. That is a DIFFERENT population — any file carrying an emit
+# shape, including gates that emit only in a non-CI mode — so it is not a
+# recount of this sentence and must not be substituted for one.
+#
+# Three emitters are deliberately NOT listed, all for one reason. The first is
+# `lint_evidence_artifacts`, for the same reason `check_rule_projection_integrity`
+# is not (see the paragraph below): it is diff-scoped, so it legitimately emits
+# `scanned: 0` on any
 # branch that adds no evidence artifact, and `min_scanned: 0` is the
 # false-count shape this file rejects. Its ledger half IS discharged — it
 # adopts `_lib/gate_ledger.ts` and `_lib/scan_scope.ts`, so
@@ -182,6 +219,28 @@ gates:
       the Taskfile parse degrades and the population silently collapses —
       cannot pass.
 
+  - id: check_rule_layer_partition
+    argv: []
+    min_scanned: 0
+    corpus: "the five project rule directories that receive a projected tree (.claude/rules, .cursor/rules, .clinerules, .windsurf/rules, .augment/rules)"
+    status: enforced
+    note: >
+      `min_scanned: 0` is the honest floor and not the false-count shape this file
+      rejects elsewhere. This gate compares a project rule directory against its
+      host's GLOBAL layer, and on a fresh CI checkout neither side exists — the
+      project trees are gitignored and no CI leg installs at user scope. So the
+      population is legitimately zero there, exactly as it is for the diff-scoped
+      gates the header paragraph carves out.
+      What keeps a zero from reading as a pass is inside the gate rather than in
+      this floor: with nothing compared it prints "nothing compared — N skipped.
+      This is not a pass" instead of a success line, and every skip names the host
+      and the unreadable path. Teeth land on a working checkout, where `task
+      preflight` runs it against real global layers.
+      It never reports a duplicate for a host whose global layer cannot be read: a
+      project copy is the only carrier when the global one is absent, so demanding
+      its removal on unread evidence would delete the rule from that host —
+      the failure `partitionEligibility`'s fail-safe ordering exists to prevent.
+
   - id: check_rule_activation_census
     argv: []
     min_scanned: 100
@@ -555,14 +614,62 @@ gates:
 
   - id: lint_plan_risk_register
     argv: []
-    min_scanned: 12
-    corpus: "ready roadmap files under agents/roadmaps/ top level (19 at baseline: 14 grandfathered + 4 draft-exempt + 1 carrying a live register)"
+    min_scanned: 1
+    corpus: "ready roadmap files under agents/roadmaps/ top level (6 as of 2026-08-21; 19 at the original baseline: 14 grandfathered + 4 draft-exempt + 1 carrying a live register)"
     status: enforced
     note: >
       Gate R1 (docs/contracts/plan-review-gates.md § 1). The count is READY
       FILES INSPECTED incl. grandfathered/draft-exempt ones — exemptions are
-      reported, never silently skipped, so a collapse below the floor means
-      the roadmaps root moved, not that the corpus went clean.
+      reported, never silently skipped.
+
+      Lowered 12 → 5 on 2026-08-21, and the SHAPE of the change matters more
+      than the number. The previous note asserted that "a collapse below the
+      floor means the roadmaps root moved, not that the corpus went clean."
+      That premise is falsified: archiving roadmaps at zero open and zero
+      deferred steps dropped the top-level count from 12 to 10 with the root
+      untouched — the corpus really did go cleaner, which is what
+      road-to-estate-drawdown exists to achieve. A floor pinned to the exact
+      corpus size fails on the drawdown succeeding, which is the wrong thing
+      to detect.
+
+      So the replacement is deliberately NOT today's count. Re-pinning at 10
+      would repeat the defect one archival later and force this note to be
+      rewritten every time a roadmap closes. 5 sits structurally below any
+      plausible active set while still catching the two failures this floor is
+      for: the roadmaps root moving, and the scanner silently reading nothing.
+      It is not a target for how many roadmaps ought to exist.
+
+      Count reconciliation, since two numbers move in this change and they
+      count different populations: this gate counts READY files at the top
+      level of agents/roadmaps/ (12 → 10 here), while
+      estate-count-budget.json::active_roadmaps counts NON-DRAFT candidates as
+      collect() sees them (11 → 10). The gap is the draft-status file each
+      population treats differently; neither number is derivable from the
+      other.
+
+      Revisit-if the active set is deliberately grown past ~15, at which point
+      5 is uninformatively low, or if a drop is ever traced to a scanner
+      defect rather than to archival — either falsifies the number, not the
+      obligation.
+
+      Lowered again 10 → 5 on 2026-08-21, during the PR drain that produced the
+      paragraph above. 10 was measured against 12 and the drain had already
+      taken the corpus to 11 — one file of headroom, so the next archival would
+      have reddened CI for the exact reason this entry says is the wrong thing to
+      detect. 5 keeps the root-moved / scanner-blind signal (both scan 0) and
+      stops the number from needing a commit per archived roadmap.
+
+      Lowered again 5 → 1, and this time the SHAPE changed rather than the
+      number. 5 was picked to have headroom and did not: the drain took the
+      corpus to 6, so a review provoked the red by archiving one roadmap. The
+      lesson is that no absolute floor survives here — this gate's sibling
+      `check_estate_count` forbids growth, so the corpus can only shrink and any
+      floor above 0 is met eventually. 1 is the smallest value that still
+      distinguishes "the scanner read something" from "the root moved or the
+      scope is dead", which is the only thing this field was ever asked to
+      detect. Revisit-if the gate gains a way to express "non-empty" directly,
+      or if a drop to 0 is ever traced to a scanner defect rather than to an
+      empty corpus.
 
   - id: check_estate_count
     argv: []
@@ -695,6 +802,18 @@ gates:
     note: >
       Emits on stderr, not stdout: `--format json` owns this gate's stdout and a
       count written there would corrupt the payload. The guard reads both streams.
+      Corpus boundary, decided 2026-08-20 (council, `leave-unscanned`): `src/skills`
+      is deliberately OUT of this gate's corpus. Pointed at it the gate reports 13
+      findings (11 `buried_iron_law`, 2 `deep_iron_law`) that nobody currently sees.
+      Extending the corpus was declined on three grounds — the gate has no warn
+      level, so a staged extension would mean adding a warning mode to an enforced
+      blocking gate; the reverted `lint_skill_top_position` attempt showed two gates
+      defining one construct incompatibly, and the disputed `deep_iron_law` class is
+      still live in those 13; and the premise for scanning skills at all
+      (post-compaction truncation keeps the file start) is unverified, the tree's one
+      measured truncation fact being ENTRY-level. Re-opening this needs a measured
+      reason plus the reported-and-baselined ratchet, not a path added to `argv`.
+      See `agents/roadmaps/archive/road-to-context-fidelity.md` § Phase 3.
 
   - id: check_cli_registry_budget_sync
     argv: []
@@ -955,3 +1074,92 @@ gates:
     # an already-listed file that GREW while the file COUNT stayed at one. That
     # last case is what pins the metric as excess lines rather than a file count;
     # a count-based ratchet passes it, and this gate must not.
+
+  # check_new_adr_evidence is DELIBERATELY NOT REGISTERED here, following this
+  # file's own two precedents rather than breaking them.
+  #
+  # It is diff-scoped: its corpus is the set of decision records ADDED on the
+  # branch (`git diff --diff-filter=A` over `docs/decisions/ADR-*.md` and
+  # `docs/adrs/<area>/NNNN-*.md`), so ZERO is the honest reading on almost every
+  # run and any floor above 0 reds every branch that adds no ADR. A
+  # `min_scanned: 0` row is the false-count shape this file's header rejects in
+  # as many words, and `lint_evidence_artifacts` and
+  # `check_rule_projection_integrity` both hit this exact wall and resolved it by
+  # staying unregistered. A row was written first and then withdrawn: the
+  # argument for it was that a registered gate at least runs under CI-identical
+  # argv, and that is real but it is not worth being the one instance that
+  # introduces the shape two neighbours were excluded to avoid. Three precedents
+  # agreeing is a convention; two agreeing and one defecting is drift.
+  #
+  # It is NOT unwatched. `task check-new-adr-evidence` and the
+  # `New accepted ADRs disclose their evidence` step in rule-backstops.yml both
+  # run it, so a broken gate fails there.
+  #
+  # And it is NOT unproven, which is the part a floor could never have supplied.
+  # This gate's normal reading is zero added records — precisely the state where
+  # a working gate and a completely broken one emit identical output
+  # (`scanned: 0`, green), so no floor separates them at any value. `--self-test`
+  # does: seven cases, four rejecting, each building a throwaway repo and driving
+  # the real CLI with `--base <sha>`, including the case an `authority_basis:
+  # owner_intent` escape must NOT buy (that field governs authority, never
+  # disclosure). Sabotage-verified — with the `## Evidence` check removed the
+  # self-test exits 1 and names the two cases that stopped firing.
+  #
+  # What is LOST, stated rather than implied: no minimum scan is held against
+  # this gate by this manifest, so a change that silently narrowed its path
+  # regexes to nothing would pass here. The self-test is what catches that, and
+  # it is a test rather than a manifest floor — it runs when someone runs it.
+
+  # adr/evidence_census is DELIBERATELY NOT REGISTERED here either, for the
+  # same structural reason and with a different cost.
+  #
+  # `matchesGatePattern` classifies on a name PREFIX
+  # (`lint|check|audit|skill|verify_`) and `evidence_census.ts` carries neither,
+  # so a row here would red `gate_population.test.ts` exactly as the
+  # `adr_cite_check` row did. The block below weighs the three fixes for that
+  # case and they apply unchanged: widening the pattern, renaming the script, or
+  # inverting the classification to be manifest-driven.
+  #
+  # What is DIFFERENT, and better than the adr_cite_check case: the census is not
+  # unwatched. `task check-adr-evidence-census` regenerates it to a temp path and
+  # diffs against the committed artifact, wired in the `ci` chain and as the
+  # `ADR evidence census up to date` step in rule-backstops.yml, so a corpus that
+  # drifted from the artifact fails there instead of here. The census emits a
+  # `scanned:` line and adopts the ledger, so what is lost is only the MINIMUM
+  # SCAN FLOOR — a change that narrowed its corpus to zero AND regenerated the
+  # artifact to match would pass both. Stated rather than implied, because the
+  # drift check is a byte-comparison and byte-agreement on an empty corpus is
+  # still agreement. Closing it properly is the same rename, in the same change
+  # that renames adr_cite_check.
+
+  # adr_cite_check is DELIBERATELY NOT REGISTERED here, and the reason is a
+  # disagreement between two gates over the same file rather than an oversight.
+  #
+  # The script is gate-shaped by behaviour: it emits one `scanned:` line, adopts
+  # the ledger, asserts its scan scope, and runs in CI as `ADR citations resolve`
+  # (rule-backstops.yml) plus `task check-adr-citations`. So it certifies
+  # something and can lie about it, which is this manifest's own criterion.
+  #
+  # But `gate_population.matchesGatePattern` classifies structurally, on a NAME
+  # PREFIX (`lint|check|audit|skill|verify_`), and `adr_cite_check.ts` carries
+  # `check` as a suffix. Registering it made `gate_population.test.ts` red:
+  # "manifest lists adr_cite_check, population does not". The two gates read the
+  # same file and disagree.
+  #
+  # Three fixes were weighed and each was worse than the gap:
+  #   - widening the pattern to a `_check$` suffix pulls in four unrelated
+  #     scripts (bench_drift_check, worktree_cleanup_check, update_check,
+  #     cmd_settings_check), forcing them into the ledger contract and reddening
+  #     the completeness ratchet for work nobody asked for;
+  #   - renaming the script to `check_adr_citations.ts` is the convention-conform
+  #     answer and touches 26 files, including archived review artefacts and a
+  #     committed diff.patch that must not be rewritten;
+  #   - manifest-driven classification is explicitly rejected by
+  #     gate_population.ts's own header (a new gate lands before anyone
+  #     registers it, so that inverts the real dependency).
+  #
+  # So the row is dropped and the floor is unregistered. What is LOST is stated
+  # rather than implied: the gate runs, but nothing holds it to a minimum scan,
+  # so a future change that silently narrowed its corpus to zero would pass. The
+  # honest close is the rename, in its own change, where 26 files of churn are
+  # the subject rather than a side effect.

package/src/config/gate-violation-baselines.json

@@ -1,66 +1,76 @@
 {
-    "$comment": "Violation ratchet for gates whose dead scan root was repaired. A repair REVEALS pre-existing debt; it does not create it. Each entry records the count revealed at repair time \u2014 the gate fails only when its count RISES. Lowering a count is a normal commit; raising one is a defect. Entries expire: a baseline that has not dropped within 56 days of `landed` (or of `reaffirmed.date`) fails its gate, because a number that never moves is suppression with extra steps. Mechanism + rationale: src/scripts/_lib/gate_baseline.ts. Decision: AI council 2026-08-02, roadmap road-to-gates-that-can-fail blocker `dead-gate-finding-triage`.",
-    "gates": {
-        "audit_user_type_axis": {
-            "count": 1,
-            "landed": "2026-08-02",
-            "note": "One orphan user-type value: `legal`. Five legal-review-prep skills (contract-review, dpa-review, nda-triage, legal-intake-triage, legal-practice-profile) declare `recommended_for_user_types: [legal]`, but there is no `user-types/legal.yml` \u2014 the pack shipped an axis value with no user-type definition behind it, and the dead scan root hid it (the audit reported used=0 orphans=0 while reading zero skills; it now reads 288 and reports used=8). Clearing this entry is a product call, not a copy edit: either author `user-types/legal.yml` (tagline, recommended packs, install-path hint) or drop `legal` from those five skills. Not decided here because both options change shipped pack behaviour."
-        },
-        "check_depth_budget": {
-            "count": 4,
-            "landed": "2026-08-06",
-            "note": "Four depth files exceed the 16,000-char per-file ceiling: rule-body-migration-inventory.md (23,782), design-antipatterns.md (20,707), roadmap-process-loop.md (20,360), orchestration-telemetry.md (16,155). Recorded rather than fixed because splitting four large reference documents is unrelated to the change that introduced the ceiling, and a gate that lands as four instant blockers is the shape this repository has already recorded as a failure. Shrink-only from here: a FIFTH over-ceiling file reds. roadmap-process-loop.md is the file that motivated the gate \u2014 it states 'Size budget: <= 4,000 chars' in its own header and is 5x over it, which is what a declared budget with nothing measuring it decays into. The ceiling is a GROWTH RATCHET, not a measured quality threshold: the 7,308-trajectory study behind the 3,500-token skill ceiling (ADR-217) was taken on procedural always-co-loaded skills, and this layer is declarative and loaded one or two files at a time. Replace the number when someone measures comprehension on THIS layer."
-        },
-        "check_gate_completeness": {
-            "count": 215,
-            "landed": "2026-08-17",
-            "note": "Ledger adoption ratchet. Gates registered under the CI task roots that carry neither per-target completeness accounting (`_lib/gate_ledger.ts`) nor a `// ledger-exempt:` marker. This is deliberately NOT a sweep: road-to-skill-ecosystem-gate-integrity Phase 1 adopts exactly three gates (lint_framework_leakage, check_references, lint_roadmap_complexity) plus the checker itself to prove the interface, and the remainder migrates as gates are touched. Shrink-only from here \u2014 the number may not grow, so a NEW gate must adopt or exempt. Lowered 216 -> 215 on 2026-08-17 by discharging the obligation three NEW arrivals had incurred, which is the standing rule being enforced rather than a drain: the registered population grew 244 -> 258 and three of the fourteen arrivals adopted nothing \u2014 `build_archive_index`, `check_gate_coverage`, `check_pr_ci_current`. The reading was 218 against this baseline of 216, and the missing arithmetic is that one PRE-existing gate (`lint_rule_skill_pack_reach`) had adopted in the meantime: 216 + 3 - 1 = 218. All three adopted a real ledger. `check_pr_ci_current` first took a reasoned exemption claiming no target population; the completion review refuted it from the file itself \u2014 the check rows ARE a population, and rows whose `state` is not a string were being dropped before any verdict ran, so a dropped row could be the failing one. The exemption was withdrawn and the ledger adopted, counting kept rows as read, dropped rows as skips, and rows a pre-row verdict never reached as out of scope. The denominator in this note was 14 stale (it said 244) and is now expressed without one, because a count pinned to a population that grows weekly rots faster than the count itself. Note what this lowering does NOT buy: headroom is zero either way, so the next un-adopted arrival reds this gate immediately \u2014 which is the intended behaviour and the reason the number is not padded. The refreshed `landed` is load-bearing rather than bookkeeping: `checkRatchet` returns `improved` BEFORE it evaluates staleness, so an entry left at 216 while the tree measures 215 would never expire at any age (pinned by tests/lib/gate_baseline.test.ts, which asserts `improved` at landed + STALE_AFTER_DAYS + 400 days) \u2014 the greenest available state and the worst, since it switches off the anti-fossilization clause exactly when someone has done real work.",
-            "falsifier": "./scripts-run src/scripts/check_gate_completeness"
-        },
-        "check_no_new_legacy_path:hardcoded-scan-roots": {
-            "count": 47,
-            "landed": "2026-08-13",
-            "note": "Lines under src/scripts/**/*.{ts,mts,mjs} that pass the retired `.agent-src.uncondensed` directory into `path.join(...)` \u2014 a filesystem path CONSTRUCTED against the tree ADR-051 abandoned, which is the exact shape of all 14 dead gates. Deliberately narrow so the number is signal, not noise: 236 raw mentions under src/scripts \u2192 147 string-literal lines \u2192 67 path-constructing lines when the pass was written, after exempting src/scripts/_lib/agent_src.ts (the shared resolver, whose job IS owning the constant \u2014 gates that route through it survived the move). Detectors holding the literal as data to match, strip, or print are not counted, and no prose, docs, tests or fixtures are read: that full-tree shape was measured at 44 files / 213 hits and rejected as the wrong lint (see the note in check_no_new_legacy_path.ts::find_offenders). Recorded at 58 rather than 67 because Phase 1's scan-root repairs were landing on this branch while the pass was being written and had already removed 9 \u2014 a baseline above the truth silently permits new violations. This count goes to ZERO as the remaining repairs land; every repair is a lowering commit. Lowered 56\u219251 by the road-to-renewal-foundation Phase 1 dead-root sweep (check_always_budget, check_portability, check_references, lint_framework_leakage, lint_agents_md). Lowered 51\u219249 by the gate-hardening adoption PR: skill_collision_clusters and skill_overlap were repointed at the live skills tree via the shared resolver \u2014 two gates that had been reading the retired container and scoring 0 of 288 skills. Lowered 49\u219248 by road-to-rule-delivery-integrity P2.2: score_skill_selection was repointed at the live skills tree via the shared resolver. It is the same shape as the two gates lowered before it \u2014 a scorer reading the retired container and scoring 0 of 288 skills \u2014 and here the consequence was sharper than a dead gate: the roadmap's pre-registered selection-accuracy criteria would have been compared against a baseline of silent zeros. Lowered 48\u219247 on 2026-08-13, and this one is a catch-up rather than a repair: the pass reported 47 against a baseline of 48, so a gain had already landed without the ratchet following it. Attributed to 455160160 (fix(evals): resolve the eval orchestrator against the live skills tree), the most recent repair in this branch's base that removes a path-constructing line \u2014 SKILLS_ROOT repointed through _lib/agent_src.ts::SRC_SKILLS(). A loose ratchet is not harmless: it silently permits one new violation to be added back without any gate noticing, which is the shape this entry exists to prevent."
-        },
-        "check_source_size_budget": {
-            "count": 19427,
-            "landed": "2026-08-20",
-            "note": "Lowered 19,570 -> 19,427 on 2026-08-20 by road-to-single-delivery Phase 5, and the -143 is a LOWERING THROUGH EXTRACTION in the sense this note's own convention requires -- reproducible with `git stash && npx tsx src/scripts/check_source_size_budget.ts`. The branch had to WIRE a feature into the two worst files here: the delivery partition (ADR-236) needs a decision point in condense.ts (2,885) and a fingerprint stamp in install.ts (5,461), and a first pass added 191 lines above the ceiling -- a straight regression this gate correctly refused. The fix was not to shrink the comments until the number fit. Three extractions carried it: the partition decision, its memo, its fail-safe ordering and its printed reason went to src/install/partitionEligibility.ts (325 lines, under the ceiling, contributing nothing here); the fingerprint itself to hostLayerFingerprint.ts; and then the one that produced the surplus -- the whole Claude `paths:` emission plan (_claude_paths_plan, _has_non_path_trigger, derive_trigger_globs, _is_unresolved_placeholder, _escape_claude_bracket, _expanded_pattern_count, the budget constant and the plan type, 169 lines) moved to src/install/claudePathsPlan.ts. That last move is not opportunistic: Phase 5.1 removed the path triggers from four package-only rules so they load unconditionally, which is a change ABOUT that plan, and the plan is host-activation-key logic that belongs beside emit_host_rules_cli.ts rather than inside the orchestrator that calls it. condense.ts ends at 2,735 -- 150 lines BELOW where the branch found it, while carrying four new call sites. Every consumer that imported the moved symbols from condense.ts still can: they are re-exported, so the extraction is invisible to callers and visible to this ratchet. Previous entry: lowered 19,626 -> 19,611 on 2026-08-17 by the release-gh-transient branch, same shape (retry loop and transient classifier extracted to _lib/gh_transient.ts, leaving release.ts a 20-line wrapper).",
-            "falsifier": "./scripts-run src/scripts/check_source_size_budget --quiet"
-        },
-        "ci-parity:local-only": {
-            "count": 165,
-            "landed": "2026-08-13",
-            "note": "Gates that `task ci` runs and no workflow does \u2014 the direction this manifest's own header calls the one that let real defects merge (16 stale index rows reached the trunk behind a local-only check-index). Revealed, not created, by repairing the checker's extraction: check_ci_local_parity built its CI-side set by regexing RAW workflow text for `task <name>` and expanding each closure, comments included. Several workflow comments contain the literal string `task ci` while stating that no workflow invokes it, so `ci` was read as an invocation and its full 247-gate closure entered the 'runs in CI' set. The count was therefore 0 BY CONSTRUCTION, and the prose documenting the gap was what suppressed it. Measured at the repair: CI-side 273 \u2192 106, undeclared_local_only 0 \u2192 167 of 247 local gates (68%). Baselined rather than hard-failed because 167 findings nobody can clear in the change that reveals them is the gate-that-lands-as-N-instant-blockers failure this repository has refused before; shrink-only keeps the property that matters, namely that a NEW gate registered in `task ci` with no workflow reds immediately. Lowering happens by wiring gates into workflows or declaring them under `local_only:` with a reason. Note the punchline, because it bounds the risk of this entry: check_ci_local_parity ITSELF runs in no workflow, so it is its own finding and this ratchet has no remote reach either \u2014 clearing the backlog is what gives it teeth. DRAINED BY ONE, 2026-08-13: check_ci_local_parity itself was wired into consistency.yml, so it left the set it exists to report. That is the whole shape of a legitimate drain here \u2014 the gate stops being local-only because a workflow now runs it, not because the declaration list grew. The floor in gate-coverage.yml moved in the same change and for the same reason: wiring two gates raised the parsed population 357 -> 360 (109 CI + 251 local). The intermediate 359 was the reading after the CI-side wiring alone and before check_gate_coverage joined the local `ci` chain; it was committed here by mistake and is corrected rather than left as a second number for the same measurement. WALKED DOWN 166 -> 165 on 2026-08-20 (road-to-session-closeout 5.2): check_ci_local_parity reported `ci-parity:local-only: 165 violation(s), baseline 166 - ratchet is loose` and asked for exactly this, so the gain cannot be silently given back. The measurement is the gate's own output on the merged tree, not a branch. What is NOT done here is the classification the same step asks for: 165 gates each need a recorded class (belongs-in-CI / developer-aid / dead), which is per-gate judgement, and migrating them would red the trunk on the landing day - the step forbids it for that reason.",
-            "falsifier": "./scripts-run src/scripts/check_ci_local_parity"
-        },
-        "gate-self-test:registered-non-adopters": {
-            "count": 24,
-            "landed": "2026-08-06",
-            "note": "Self-test adoption ratchet, expressed as a shrink-only NON-adopter count. 24 of the 32 gates registered enforced in gate-coverage.yml carry neither a `--self-test` (`_lib/gate_self_test.ts`) nor a `// self-test-exempt: <reason>` marker, so nothing proves they DISCRIMINATE \u2014 an enforced `scanned:` floor proves a gate read something, only a self-test proves the reading changes the verdict. Deliberately NOT an adoption column ratcheted upward: check_gate_coverage.ts's own report_hardening_ratchet rejects that shape in this same file (\"grades the solution instead of the problem\"), and the sibling adoption ratchet (check_gate_completeness) is already a non-adopter count. Population is the manifest, not every script under src/scripts: padding the manifest to move this number would mean adding a floor per row. Shrink-only from here \u2014 a NEW registered gate must adopt or exempt. Migration mechanism, same as the ledger ratchet above: gates adopt as they are touched, not in a sweep \u2014 the eight current adopters were each added alongside a real change to that gate. The 56-day non-stagnation clause applies unchanged, so a count parked here fails on its own.",
-            "falsifier": "./scripts-run src/scripts/check_gate_coverage"
-        },
-        "lint_handoffs": {
-            "count": 18,
-            "landed": "2026-08-02",
-            "note": "All 18 are `handoff_tier_mismatch`: a senior-tier skill links to a peer whose frontmatter carries no `tier` (reported as tier='unset') \u2014 e.g. build-buy-partner \u2192 adr-create, customer-research \u2192 po-discovery, gated-reach \u2192 threat-modeling. This is a tier-metadata backfill backlog on the LINKED-TO skills, not a broken link. The one genuinely dangling link revealed by the same repair (competitive-positioning \u2192 a non-existent ../analyze-reference-repo/SKILL.md; it is a command, not a skill) was FIXED in the repairing commit rather than baselined. Clearing this entry means backfilling `tier` on the linked-to skills, or deciding that `tier: unset` on a link target is not a violation and narrowing the rule."
-        },
-        "lint_pack_boundaries": {
-            "count": 337,
-            "landed": "2026-08-02",
-            "note": "Markdown links between source artefacts that cross a pack boundary the source pack's `requires` graph does not reach \u2014 revealed when the scan root moved off the deleted packages/ tree to the live corpus (0 -> 753 artefacts scanned). 82 distinct pack pairs across 131 source files; largest single class is engineering-base -> meta (73), skills in a consumer-installable pack citing this package's own maintenance skills. Each is a link that dangles for a consumer who installs the source pack without the target's. Judged real drift rather than an over-strict gate because it is long-tailed \u2014 top bucket 22%, no single systematic cause \u2014 unlike the lint_artefact_frontmatter case, where one required shape had 0-of-288 adoption. Fix per finding: add the edge to packs.yml requires, retarget the link, or move the artefact."
-        },
-        "lint_roadmap_blockers:decidability": {
-            "count": 12,
-            "landed": "2026-08-21",
-            "note": "Open blocker entries that are not yet DECIDABLE by the person who owns them \u2014 missing `Recommendation:` (which option and why), missing `If you do nothing:` (the cost of the non-decision), or a `What to do:` carrying no command, path or enumerated option set. Revealed by adding the decidability half to lint_roadmap_blockers, not created by it: the five original fields describe a situation and never made it actionable, so a blocker could pass the contract while handing the analysis back to the person least equipped to redo it. Measured at introduction: 36 open entries, ALL 36 undecidable, and 14 of the 46 total entries carry no command/path/option at all in `What to do`. Baselined rather than hard-failed because 36 findings nobody can clear in the change that reveals them is the gate-that-lands-as-N-instant-blockers failure this repository has refused before \u2014 the same reasoning as ci-parity:local-only above. First measured at 34 rather than 36 because the two blockers on the branch that introduced this (`ac3-false-positive-reading`, `publish-side-marker-invariant`) were rewritten to the new standard and then closed in the same change; a baseline above the truth silently permits new violations. Re-measured at 38 when the branch was re-pointed at current main, and the +4 is population growth rather than a loosening: main gained the four `road-to-inbox-harvest-2026-08-d-*` roadmaps (context-ledger, runtime-skill-routing, scheduled-deprecation, top-band-model-economy), each carrying one open blocker authored before this requirement existed. That is the honest number for the population this gate now scans \u2014 recording 34 against a 38-entry tree would have failed the gate on its own introducing commit and taught the next reader to distrust it. Lowering happens one blocker at a time, and every roadmap touched for other reasons is a chance to lower it. Lowered 38\u219237 by the scheduled-deprecation PR, which is exactly that shape: code-graph-removal-authorisation was rewritten to the new standard because the roadmap was being touched anyway. Its `If you do nothing:` was one of the few that was not the common no-cost answer \u2014 the release gate that PR ships would have refused the next major cut while the blocker stayed open. Recorded in the past tense because the maintainer then RESOLVED that blocker on the trunk by withdrawing the commitment, so nothing refuses a cut today. The lowering stands on the rewrite, not on the consequence, and the sentence is corrected rather than deleted because a baseline note arguing from a retracted state is how a ratchet's rationale rots. The 56-day expiry is the point, not a hazard: if this number has not moved in two months, the fields are not being written and the gate should fail until someone decides whether the requirement is real. Lowered 37\u219234 when the scheduled-deprecation branch merged current main: three blockers were resolved or rewritten on the trunk in the meantime, so the drop is a genuine drain plus a merge, not a loosening. Re-tightened here rather than left loose because a ratchet that lags its own tree permits violations back in silently. Lowered 34\u219229 on 2026-08-16 by rewriting six entries to the new standard (raw-capture-needs-host-env, telemetry-sample-size, repo-admin-and-usage, benchmark-spend, measurement-a-no-per-arm-builder-tier, measurement-b-no-renderable-lane-pair) in a PR whose only purpose was the drain \u2014 the first lowering not taken on the back of another change, which the note above says is the usual shape. Two findings from that pass are worth keeping. The tree measured 35 against a baseline of 34 BEFORE the drain, i.e. the ratchet was already red on main and nobody knew: this gate runs in `task ci`, no workflow invokes `task ci`, and the local run that would have shown it was read through a `tail`, which hid every row above the last six. And four of the six `If you do nothing:` answers are honestly that nothing degrades \u2014 that is the true answer for a held measurement, and writing it is what makes the entry decidable, so the field earns its place by being cheap to answer truthfully rather than by always carrying a cost. Lowered 29 \u2192 27 on 2026-08-16 by the merge of this branch and the trunk drain above, and the arithmetic matters more than the delta: both sides rewrote entries from the same tree of 35 independently \u2014 the trunk six (recorded above), this branch two (deferred-finding-decision-reopen, spent-inbox-artifacts-await-deletion, rewritten because road-to-inbox-harvest-residuals was being touched anyway). The sets are disjoint, so the merged tree measures 27 and the gate itself reported the slack. Tightened rather than banked: 29 against a 27-entry tree would absorb the next two undecidable blockers without anything noticing, which is the lag this note has now corrected twice. Lowered 27 \u2192 26 on 2026-08-16 by rewriting `picktier-wire-or-archive` (road-to-inbox-harvest-2026-08-d-top-band-model-economy), and this one is the usual shape the note describes rather than the dedicated-drain shape: the roadmap was being worked for its own Step 4.1 and the entry was rewritten because it was open on the desk. Worth recording for what writing the two fields SURFACED, which is the argument for the requirement: `Recommendation:` forced the question 'which option, and why' and the honest answer turned out not to be one of the three options at all \u2014 the entry defers to a converged council verdict, and the council cannot reach quorum (re-measured 0 of 2 the same day), so the actionable next step is a one-line fix to a user-global config file outside this repository. Five fields had described that situation accurately for two weeks without anyone noticing that the named blocker was infrastructure rather than judgement. Lowered 26 \u2192 22 on 2026-08-18 by road-to-estate-drawdown Step 0.1, which needed the two fields for a different reason and therefore drained six entries as a by-product: the consolidated decision sheet renders a DEFAULT plus that default's provenance per user-owned decision, and 7 of the 21 live entries recorded no `Recommendation:` to derive one from. Six were rewritten to the new standard (skill-activation-window, b-convergence-machine, maintainer-blind-ratings, real-orchestration-usage, human-gated-live-trigger-eval, manual-rubric-rater), each marked `agent-drafted` in the field itself so the sheet can label the provenance rather than presenting a drafted default beside a maintainer-recorded one as the same kind of statement. The seventh, `road-to-gated-reach-followup`'s legacy `> Blocked until \u2026` note, is NOT counted here and was not rewritten: it is a blocked-until note rather than a `### blocker:` entry, so it has no field to carry, and converting it is a roadmap restructure this change did not take. **The tree was RED at 28 against this baseline of 26 before the drain, and again nobody knew** \u2014 the same failure the 34\u219229 paragraph above records, one gate cycle later: two entries had been added on the trunk without the fields, and this gate runs in `task ci`, which no workflow invokes. That is the second recorded instance of this ratchet being red on main undetected, which is an argument about where it runs rather than about its number. `human-gated-live-trigger-eval` also gained a command in `What to do:` (`rule_trigger_eval`, verified to exist at src/scripts/) because the field carried a path-less instruction; the other five needed only the two decidability fields. Lowered 22 -> 20 on 2026-08-19 by road-to-estate-drawdown Phase 2 triage batch 1, and RE-MEASURED after two merges of origin/main rather than carried over: the trunk gained roadmaps in between (the long-horizon pair and the council-fallback set), and the live count over the merged tree is still 20 across 39 roadmaps. Six roadmaps moved to later/, and two of them carried an undecidable open blocker each. The gain is a side effect of parking rather than of drafting a recommendation, so it is recorded as such \u2014 the two entries are unchanged and still undecidable where they sit; they are simply outside the active tree this ratchet scans. Walked down because the gate reported itself loose, and a baseline above the measurement is headroom a later change can spend with nothing objecting. Lowered 20 -> 17 on 2026-08-20 by the road-to-frontend-skill-application closeout, and the shape is the usual one this note describes: the roadmap was being closed for its own reasons and its three open blockers were on the desk anyway. Two (ui-corpus-has-no-ui, enforcement-evidence) left the population by being RESOLVED rather than rewritten, which is the cleanest way to lower this number and the only one that does not add prose; the third (ui-session-capture-window) was rewritten to the new standard. Worth recording what writing the two fields surfaced, since that is the standing argument for the requirement: on ui-corpus-has-no-ui, `Recommendation:` forced the question \"which option, and why\" and the honest answer was that the entry's own premise was false \u2014 it asserted the question could not be answered without a human naming a consumer store, and the stores were already on the machine, readable by the command the entry itself printed. Five fields had described that situation accurately for a week without anyone noticing the blocker was a measurement nobody had run. Tightened rather than banked because 20 against a 17-entry tree is headroom the next three undecidable blockers would spend with nothing objecting, which is the lag the paragraphs above have now corrected three times. Walked 17 -> 12 at the drain/road-to-frontend-skill-application merge, 2026-08-21, on the gate's own loose-ratchet report. Re-measured against the merged tree rather than attributed: sibling drain PRs made entries decidable in parallel, so this branch's share is not separable here and is not claimed. Banked at the measurement rather than left as headroom a later change could spend.",
-            "falsifier": "./scripts-run src/scripts/lint_roadmap_blockers --quiet"
-        },
-        "lint_settings_classes:derivable-surface": {
-            "count": 83,
-            "landed": "2026-08-12",
-            "note": "Anti-regrowth ratchet for the settings disposition axis. 83 of 134 template leaves are classified `derivable` \u2014 the mechanism can decide from the situation, so the key should not exist \u2014 and each names the replacement that would decide instead (`lint_settings_classes` rejects the bare label, which is what keeps the class from becoming a synonym for \"inconvenient to keep\"). This is a DELETION QUEUE, not 83 deletions: a row stays until the mechanism it names actually exists, and deleting a key ahead of its mechanism is a silently-changed default rather than a simplification. Shrink-only from here, which is the anti-regrowth half \u2014 a NEW key classified `derivable` raises the number and reds, so a key whose answer the situation already carries cannot be added at all. The 56-day non-stagnation clause is the other half and is wanted here rather than tolerated: a queue that never drains is a stated direction nobody is walking. The queue opened at 88 of 140 on 2026-08-12 and drained to 83 of 134 the same day: six keys had no reader whatsoever (telegraph.speak_scope, chat_history.max_size_kb, chat_history.on_overflow, quality.wait_for_remote_ci, screenshots.data_bearing_gate, legal_review_prep.consented_at), five of them `derivable` and one a `consent` row reclassified because a published-egress confirmation is a Hard Floor no setting may lift. An unread key is the only deletion that cannot silently change a default, which is why that batch went first. The classification itself is docs/contracts/settings-classes.md.",
-            "falsifier": "./scripts-run src/scripts/lint_settings_classes"
-        }
+  "$comment": "Violation ratchet for gates whose dead scan root was repaired. A repair REVEALS pre-existing debt; it does not create it. Each entry records the count revealed at repair time \u2014 the gate fails only when its count RISES. Lowering a count is a normal commit; raising one is a defect. Entries expire: a baseline that has not dropped within 56 days of `landed` (or of `reaffirmed.date`) fails its gate, because a number that never moves is suppression with extra steps. Mechanism + rationale: src/scripts/_lib/gate_baseline.ts. Decision: AI council 2026-08-02, roadmap road-to-gates-that-can-fail blocker `dead-gate-finding-triage`.",
+  "gates": {
+    "audit_user_type_axis": {
+      "count": 1,
+      "landed": "2026-08-02",
+      "note": "One orphan user-type value: `legal`. Five legal-review-prep skills (contract-review, dpa-review, nda-triage, legal-intake-triage, legal-practice-profile) declare `recommended_for_user_types: [legal]`, but there is no `user-types/legal.yml` \u2014 the pack shipped an axis value with no user-type definition behind it, and the dead scan root hid it (the audit reported used=0 orphans=0 while reading zero skills; it now reads 288 and reports used=8). Clearing this entry is a product call, not a copy edit: either author `user-types/legal.yml` (tagline, recommended packs, install-path hint) or drop `legal` from those five skills. Not decided here because both options change shipped pack behaviour."
+    },
+    "check_depth_budget": {
+      "count": 4,
+      "landed": "2026-08-06",
+      "note": "Four depth files exceed the 16,000-char per-file ceiling: rule-body-migration-inventory.md (23,782), design-antipatterns.md (20,707), roadmap-process-loop.md (20,360), orchestration-telemetry.md (16,155). Recorded rather than fixed because splitting four large reference documents is unrelated to the change that introduced the ceiling, and a gate that lands as four instant blockers is the shape this repository has already recorded as a failure. Shrink-only from here: a FIFTH over-ceiling file reds. roadmap-process-loop.md is the file that motivated the gate \u2014 it states 'Size budget: <= 4,000 chars' in its own header and is 5x over it, which is what a declared budget with nothing measuring it decays into. The ceiling is a GROWTH RATCHET, not a measured quality threshold: the 7,308-trajectory study behind the 3,500-token skill ceiling (ADR-217) was taken on procedural always-co-loaded skills, and this layer is declarative and loaded one or two files at a time. Replace the number when someone measures comprehension on THIS layer."
+    },
+    "check_gate_completeness": {
+      "count": 215,
+      "landed": "2026-08-17",
+      "note": "Ledger adoption ratchet. Gates registered under the CI task roots that carry neither per-target completeness accounting (`_lib/gate_ledger.ts`) nor a `// ledger-exempt:` marker. This is deliberately NOT a sweep: road-to-skill-ecosystem-gate-integrity Phase 1 adopts exactly three gates (lint_framework_leakage, check_references, lint_roadmap_complexity) plus the checker itself to prove the interface, and the remainder migrates as gates are touched. Shrink-only from here \u2014 the number may not grow, so a NEW gate must adopt or exempt. Lowered 216 -> 215 on 2026-08-17 by discharging the obligation three NEW arrivals had incurred, which is the standing rule being enforced rather than a drain: the registered population grew 244 -> 258 and three of the fourteen arrivals adopted nothing \u2014 `build_archive_index`, `check_gate_coverage`, `check_pr_ci_current`. The reading was 218 against this baseline of 216, and the missing arithmetic is that one PRE-existing gate (`lint_rule_skill_pack_reach`) had adopted in the meantime: 216 + 3 - 1 = 218. All three adopted a real ledger. `check_pr_ci_current` first took a reasoned exemption claiming no target population; the completion review refuted it from the file itself \u2014 the check rows ARE a population, and rows whose `state` is not a string were being dropped before any verdict ran, so a dropped row could be the failing one. The exemption was withdrawn and the ledger adopted, counting kept rows as read, dropped rows as skips, and rows a pre-row verdict never reached as out of scope. The denominator in this note was 14 stale (it said 244) and is now expressed without one, because a count pinned to a population that grows weekly rots faster than the count itself. Note what this lowering does NOT buy: headroom is zero either way, so the next un-adopted arrival reds this gate immediately \u2014 which is the intended behaviour and the reason the number is not padded. The refreshed `landed` is load-bearing rather than bookkeeping: `checkRatchet` returns `improved` BEFORE it evaluates staleness, so an entry left at 216 while the tree measures 215 would never expire at any age (pinned by tests/lib/gate_baseline.test.ts, which asserts `improved` at landed + STALE_AFTER_DAYS + 400 days) \u2014 the greenest available state and the worst, since it switches off the anti-fossilization clause exactly when someone has done real work.",
+      "falsifier": "./scripts-run src/scripts/check_gate_completeness"
+    },
+    "check_no_new_legacy_path:hardcoded-scan-roots": {
+      "count": 47,
+      "landed": "2026-08-13",
+      "note": "Lines under src/scripts/**/*.{ts,mts,mjs} that pass the retired `.agent-src.uncondensed` directory into `path.join(...)` \u2014 a filesystem path CONSTRUCTED against the tree ADR-051 abandoned, which is the exact shape of all 14 dead gates. Deliberately narrow so the number is signal, not noise: 236 raw mentions under src/scripts \u2192 147 string-literal lines \u2192 67 path-constructing lines when the pass was written, after exempting src/scripts/_lib/agent_src.ts (the shared resolver, whose job IS owning the constant \u2014 gates that route through it survived the move). Detectors holding the literal as data to match, strip, or print are not counted, and no prose, docs, tests or fixtures are read: that full-tree shape was measured at 44 files / 213 hits and rejected as the wrong lint (see the note in check_no_new_legacy_path.ts::find_offenders). Recorded at 58 rather than 67 because Phase 1's scan-root repairs were landing on this branch while the pass was being written and had already removed 9 \u2014 a baseline above the truth silently permits new violations. This count goes to ZERO as the remaining repairs land; every repair is a lowering commit. Lowered 56\u219251 by the road-to-renewal-foundation Phase 1 dead-root sweep (check_always_budget, check_portability, check_references, lint_framework_leakage, lint_agents_md). Lowered 51\u219249 by the gate-hardening adoption PR: skill_collision_clusters and skill_overlap were repointed at the live skills tree via the shared resolver \u2014 two gates that had been reading the retired container and scoring 0 of 288 skills. Lowered 49\u219248 by road-to-rule-delivery-integrity P2.2: score_skill_selection was repointed at the live skills tree via the shared resolver. It is the same shape as the two gates lowered before it \u2014 a scorer reading the retired container and scoring 0 of 288 skills \u2014 and here the consequence was sharper than a dead gate: the roadmap's pre-registered selection-accuracy criteria would have been compared against a baseline of silent zeros. Lowered 48\u219247 on 2026-08-13, and this one is a catch-up rather than a repair: the pass reported 47 against a baseline of 48, so a gain had already landed without the ratchet following it. Attributed to 455160160 (fix(evals): resolve the eval orchestrator against the live skills tree), the most recent repair in this branch's base that removes a path-constructing line \u2014 SKILLS_ROOT repointed through _lib/agent_src.ts::SRC_SKILLS(). A loose ratchet is not harmless: it silently permits one new violation to be added back without any gate noticing, which is the shape this entry exists to prevent."
+    },
+    "check_source_size_budget": {
+      "count": 19397,
+      "landed": "2026-08-22",
+      "note": "Lowered 19,427 -> 19,397 on 2026-08-22 by road-to-single-rule-layer Phase 4, and the -30 is again a LOWERING THROUGH EXTRACTION rather than a comment trim -- the same shape and for the same reason as the previous entry. The branch had to wire a per-host partition decision into condense.ts (2,735), which is the worst file here, and a first pass added 105 lines above the ceiling: this gate correctly refused it. Two extractions carried it. The partition decision itself -- the fail-safe ordering, the measurement it was derived from, and why the classification source is dist/ rather than src/ -- went to src/install/ruleLayerPartition.ts, leaving a 10-line adapter behind. Then the older, weaker duplicate-avoidance mechanism moved with it: USER_SCOPE_RULE_DIRS and _dedupable_rules (projection.scope_dedup, default off, byte-identity keyed) are the same subject as the partition, and having one of the two in a 2,700-line generator and the other in src/install/ is how a reader ends up reading only one. Behaviour is unchanged there; only the MODULE_STATE reads became parameters. condense.ts ends at 2,705 -- 30 lines below where this branch found it, while carrying four new call sites for the partition and the .windsurfrules legacy surface. Lowered 19,570 -> 19,427 on 2026-08-20 by road-to-single-delivery Phase 5, and the -143 is a LOWERING THROUGH EXTRACTION in the sense this note's own convention requires -- reproducible with `git stash && npx tsx src/scripts/check_source_size_budget.ts`. The branch had to WIRE a feature into the two worst files here: the delivery partition (ADR-236) needs a decision point in condense.ts (2,885) and a fingerprint stamp in install.ts (5,461), and a first pass added 191 lines above the ceiling -- a straight regression this gate correctly refused. The fix was not to shrink the comments until the number fit. Three extractions carried it: the partition decision, its memo, its fail-safe ordering and its printed reason went to src/install/partitionEligibility.ts (325 lines, under the ceiling, contributing nothing here); the fingerprint itself to hostLayerFingerprint.ts; and then the one that produced the surplus -- the whole Claude `paths:` emission plan (_claude_paths_plan, _has_non_path_trigger, derive_trigger_globs, _is_unresolved_placeholder, _escape_claude_bracket, _expanded_pattern_count, the budget constant and the plan type, 169 lines) moved to src/install/claudePathsPlan.ts. That last move is not opportunistic: Phase 5.1 removed the path triggers from four package-only rules so they load unconditionally, which is a change ABOUT that plan, and the plan is host-activation-key logic that belongs beside emit_host_rules_cli.ts rather than inside the orchestrator that calls it. condense.ts ends at 2,735 -- 150 lines BELOW where the branch found it, while carrying four new call sites. Every consumer that imported the moved symbols from condense.ts still can: they are re-exported, so the extraction is invisible to callers and visible to this ratchet. Previous entry: lowered 19,626 -> 19,611 on 2026-08-17 by the release-gh-transient branch, same shape (retry loop and transient classifier extracted to _lib/gh_transient.ts, leaving release.ts a 20-line wrapper).",
+      "falsifier": "./scripts-run src/scripts/check_source_size_budget --quiet"
+    },
+    "ci-parity:local-only": {
+      "count": 165,
+      "landed": "2026-08-13",
+      "note": "Gates that `task ci` runs and no workflow does \u2014 the direction this manifest's own header calls the one that let real defects merge (16 stale index rows reached the trunk behind a local-only check-index). Revealed, not created, by repairing the checker's extraction: check_ci_local_parity built its CI-side set by regexing RAW workflow text for `task <name>` and expanding each closure, comments included. Several workflow comments contain the literal string `task ci` while stating that no workflow invokes it, so `ci` was read as an invocation and its full 247-gate closure entered the 'runs in CI' set. The count was therefore 0 BY CONSTRUCTION, and the prose documenting the gap was what suppressed it. Measured at the repair: CI-side 273 \u2192 106, undeclared_local_only 0 \u2192 167 of 247 local gates (68%). Baselined rather than hard-failed because 167 findings nobody can clear in the change that reveals them is the gate-that-lands-as-N-instant-blockers failure this repository has refused before; shrink-only keeps the property that matters, namely that a NEW gate registered in `task ci` with no workflow reds immediately. Lowering happens by wiring gates into workflows or declaring them under `local_only:` with a reason. Note the punchline, because it bounds the risk of this entry: check_ci_local_parity ITSELF runs in no workflow, so it is its own finding and this ratchet has no remote reach either \u2014 clearing the backlog is what gives it teeth. DRAINED BY ONE, 2026-08-13: check_ci_local_parity itself was wired into consistency.yml, so it left the set it exists to report. That is the whole shape of a legitimate drain here \u2014 the gate stops being local-only because a workflow now runs it, not because the declaration list grew. The floor in gate-coverage.yml moved in the same change and for the same reason: wiring two gates raised the parsed population 357 -> 360 (109 CI + 251 local). The intermediate 359 was the reading after the CI-side wiring alone and before check_gate_coverage joined the local `ci` chain; it was committed here by mistake and is corrected rather than left as a second number for the same measurement. WALKED DOWN 166 -> 165 on 2026-08-20 (road-to-session-closeout 5.2): check_ci_local_parity reported `ci-parity:local-only: 165 violation(s), baseline 166 - ratchet is loose` and asked for exactly this, so the gain cannot be silently given back. The measurement is the gate's own output on the merged tree, not a branch. What is NOT done here is the classification the same step asks for: 165 gates each need a recorded class (belongs-in-CI / developer-aid / dead), which is per-gate judgement, and migrating them would red the trunk on the landing day - the step forbids it for that reason.",
+      "falsifier": "./scripts-run src/scripts/check_ci_local_parity"
+    },
+    "gate-self-test:registered-non-adopters": {
+      "count": 24,
+      "landed": "2026-08-06",
+      "note": "Self-test adoption ratchet, expressed as a shrink-only NON-adopter count. 24 of the 32 gates registered enforced in gate-coverage.yml carry neither a `--self-test` (`_lib/gate_self_test.ts`) nor a `// self-test-exempt: <reason>` marker, so nothing proves they DISCRIMINATE \u2014 an enforced `scanned:` floor proves a gate read something, only a self-test proves the reading changes the verdict. Deliberately NOT an adoption column ratcheted upward: check_gate_coverage.ts's own report_hardening_ratchet rejects that shape in this same file (\"grades the solution instead of the problem\"), and the sibling adoption ratchet (check_gate_completeness) is already a non-adopter count. Population is the manifest, not every script under src/scripts: padding the manifest to move this number would mean adding a floor per row. Shrink-only from here \u2014 a NEW registered gate must adopt or exempt. Migration mechanism, same as the ledger ratchet above: gates adopt as they are touched, not in a sweep \u2014 the eight current adopters were each added alongside a real change to that gate. The 56-day non-stagnation clause applies unchanged, so a count parked here fails on its own.",
+      "falsifier": "./scripts-run src/scripts/check_gate_coverage"
+    },
+    "lint_handoffs": {
+      "count": 18,
+      "landed": "2026-08-02",
+      "note": "All 18 are `handoff_tier_mismatch`: a senior-tier skill links to a peer whose frontmatter carries no `tier` (reported as tier='unset') \u2014 e.g. build-buy-partner \u2192 adr-create, customer-research \u2192 po-discovery, gated-reach \u2192 threat-modeling. This is a tier-metadata backfill backlog on the LINKED-TO skills, not a broken link. The one genuinely dangling link revealed by the same repair (competitive-positioning \u2192 a non-existent ../analyze-reference-repo/SKILL.md; it is a command, not a skill) was FIXED in the repairing commit rather than baselined. Clearing this entry means backfilling `tier` on the linked-to skills, or deciding that `tier: unset` on a link target is not a violation and narrowing the rule."
+    },
+    "lint_pack_boundaries": {
+      "count": 337,
+      "landed": "2026-08-02",
+      "note": "Markdown links between source artefacts that cross a pack boundary the source pack's `requires` graph does not reach \u2014 revealed when the scan root moved off the deleted packages/ tree to the live corpus (0 -> 753 artefacts scanned). 82 distinct pack pairs across 131 source files; largest single class is engineering-base -> meta (73), skills in a consumer-installable pack citing this package's own maintenance skills. Each is a link that dangles for a consumer who installs the source pack without the target's. Judged real drift rather than an over-strict gate because it is long-tailed \u2014 top bucket 22%, no single systematic cause \u2014 unlike the lint_artefact_frontmatter case, where one required shape had 0-of-288 adoption. Fix per finding: add the edge to packs.yml requires, retarget the link, or move the artefact."
+    },
+    "lint_provenance_vocabulary:permanence-language": {
+      "count": 11,
+      "landed": "2026-08-21",
+      "note": "Permanence language in an ADR record's load-bearing positions -- its `# ` title, its `decision:` frontmatter slug, its `## Decision` section, or an `Addendum` heading. Revealed by adding rule 4 to lint_provenance_vocabulary, not created by it: ADR-240 section 6 states the doctrine (`forever`, `permanently`, `never revisit`, `never reconsider` and their kin do not belong in a mechanism decision's title, slug or Decision section) and the corpus had five records violating it on the day the rule landed. Baselined rather than hard-failed because a gate that lands as five instant blockers is the shape this repository has already refused (see ci-parity:local-only and lint_roadmap_blockers:decidability above), and every one of the five is a content decision rather than a copy edit: ADR-208 (slug `dist-agent-src-keep-forever`, title \"kept forever\", Decision \"KEEP -- permanently\") is the standing self-contradiction ADR-240 cites, since lines 10-11 of its own frontmatter carry the conditions under which it should be reopened; ADR-107:37 and ADR-108:5/12/24/29 are owner PURPOSE statements (\"open-source forever\", \"now or in the future\", \"liability disclaimer is permanent\") whose fix is the escape ADR-240 names -- record `authority_basis: owner_intent` -- which is an owner ruling, not an agent edit; ADR-122:77/154 and ADR-124:229 assert a permanently-off default whose honest form is a default-off setting plus a reopen condition. The rule is deliberately SCOPED and the scoping is the gate: the same word list applied to whole files matches many more, and the large majority are legitimate. The comparison figure is given with its method because 'unscoped' has two readings and an earlier version of this note picked one without saying which, giving 41 lines across 30 records which reproduces neither way: the bare word list over both ADR surfaces matches 45 lines in 26 records, and the word list PLUS the lint's own line filters (code fences, quoted spans, table rows, in-sentence negation, attributive `permanent`, hyphen compounds) but WITHOUT position scoping matches 29 in 17. Scoping to the four load-bearing positions takes it to 11 in 5. The 29-to-11 step is what the scoping buys; the 45 is what a naive grep would report -- an `## Alternatives considered` entry rejecting a permanence option (ADR-105:97, ADR-137:148-151, ADR-212:133), a `## Consequences` line reporting one (ADR-136:90, ADR-208:73, ADR-216:278/298), a `## Context` recounting one (ADR-042:36, ADR-232:34), a quoted open question (ADR-201:342/343) and a table rationale cell bounding a failure mode (ADR-074:41, ADR-225:64) are all describing a permanence claim rather than making one. Shrink-only from here, which is the property that matters: a NEW record asserting permanence in any of the four positions reds this gate immediately, and headroom is zero. Lowering happens one record at a time -- convert an owner purpose statement to `authority_basis: owner_intent`, or restate a mechanism decision with the reopen condition it already carries. The 56-day expiry is the point rather than a hazard: if this number has not moved in two months, the doctrine is not being applied and the gate should fail until someone decides whether it is real.",
+      "falsifier": "./scripts-run src/scripts/lint_provenance_vocabulary"
+    },
+    "lint_settings_classes:derivable-surface": {
+      "count": 83,
+      "landed": "2026-08-12",
+      "note": "Anti-regrowth ratchet for the settings disposition axis. 83 of 134 template leaves are classified `derivable` \u2014 the mechanism can decide from the situation, so the key should not exist \u2014 and each names the replacement that would decide instead (`lint_settings_classes` rejects the bare label, which is what keeps the class from becoming a synonym for \"inconvenient to keep\"). This is a DELETION QUEUE, not 83 deletions: a row stays until the mechanism it names actually exists, and deleting a key ahead of its mechanism is a silently-changed default rather than a simplification. Shrink-only from here, which is the anti-regrowth half \u2014 a NEW key classified `derivable` raises the number and reds, so a key whose answer the situation already carries cannot be added at all. The 56-day non-stagnation clause is the other half and is wanted here rather than tolerated: a queue that never drains is a stated direction nobody is walking. The queue opened at 88 of 140 on 2026-08-12 and drained to 83 of 134 the same day: six keys had no reader whatsoever (telegraph.speak_scope, chat_history.max_size_kb, chat_history.on_overflow, quality.wait_for_remote_ci, screenshots.data_bearing_gate, legal_review_prep.consented_at), five of them `derivable` and one a `consent` row reclassified because a published-egress confirmation is a Hard Floor no setting may lift. An unread key is the only deletion that cannot silently change a default, which is why that batch went first. The classification itself is docs/contracts/settings-classes.md.",
+      "falsifier": "./scripts-run src/scripts/lint_settings_classes"
     }
+  },
+  "retired": {
+    "lint_roadmap_blockers:decidability": {
+      "count": 0,
+      "landed": "2026-08-22",
+      "note": "Open blocker entries that are not yet DECIDABLE by the person who owns them \u2014 missing `Recommendation:` (which option and why), missing `If you do nothing:` (the cost of the non-decision), or a `What to do:` carrying no command, path or enumerated option set. Revealed by adding the decidability half to lint_roadmap_blockers, not created by it: the five original fields describe a situation and never made it actionable, so a blocker could pass the contract while handing the analysis back to the person least equipped to redo it. Measured at introduction: 36 open entries, ALL 36 undecidable, and 14 of the 46 total entries carry no command/path/option at all in `What to do`. Baselined rather than hard-failed because 36 findings nobody can clear in the change that reveals them is the gate-that-lands-as-N-instant-blockers failure this repository has refused before \u2014 the same reasoning as ci-parity:local-only above. First measured at 34 rather than 36 because the two blockers on the branch that introduced this (`ac3-false-positive-reading`, `publish-side-marker-invariant`) were rewritten to the new standard and then closed in the same change; a baseline above the truth silently permits new violations. Re-measured at 38 when the branch was re-pointed at current main, and the +4 is population growth rather than a loosening: main gained the four `road-to-inbox-harvest-2026-08-d-*` roadmaps (context-ledger, runtime-skill-routing, scheduled-deprecation, top-band-model-economy), each carrying one open blocker authored before this requirement existed. That is the honest number for the population this gate now scans \u2014 recording 34 against a 38-entry tree would have failed the gate on its own introducing commit and taught the next reader to distrust it. Lowering happens one blocker at a time, and every roadmap touched for other reasons is a chance to lower it. Lowered 38\u219237 by the scheduled-deprecation PR, which is exactly that shape: code-graph-removal-authorisation was rewritten to the new standard because the roadmap was being touched anyway. Its `If you do nothing:` was one of the few that was not the common no-cost answer \u2014 the release gate that PR ships would have refused the next major cut while the blocker stayed open. Recorded in the past tense because the maintainer then RESOLVED that blocker on the trunk by withdrawing the commitment, so nothing refuses a cut today. The lowering stands on the rewrite, not on the consequence, and the sentence is corrected rather than deleted because a baseline note arguing from a retracted state is how a ratchet's rationale rots. The 56-day expiry is the point, not a hazard: if this number has not moved in two months, the fields are not being written and the gate should fail until someone decides whether the requirement is real. Lowered 37\u219234 when the scheduled-deprecation branch merged current main: three blockers were resolved or rewritten on the trunk in the meantime, so the drop is a genuine drain plus a merge, not a loosening. Re-tightened here rather than left loose because a ratchet that lags its own tree permits violations back in silently. Lowered 34\u219229 on 2026-08-16 by rewriting six entries to the new standard (raw-capture-needs-host-env, telemetry-sample-size, repo-admin-and-usage, benchmark-spend, measurement-a-no-per-arm-builder-tier, measurement-b-no-renderable-lane-pair) in a PR whose only purpose was the drain \u2014 the first lowering not taken on the back of another change, which the note above says is the usual shape. Two findings from that pass are worth keeping. The tree measured 35 against a baseline of 34 BEFORE the drain, i.e. the ratchet was already red on main and nobody knew: this gate runs in `task ci`, no workflow invokes `task ci`, and the local run that would have shown it was read through a `tail`, which hid every row above the last six. And four of the six `If you do nothing:` answers are honestly that nothing degrades \u2014 that is the true answer for a held measurement, and writing it is what makes the entry decidable, so the field earns its place by being cheap to answer truthfully rather than by always carrying a cost. Lowered 29 \u2192 27 on 2026-08-16 by the merge of this branch and the trunk drain above, and the arithmetic matters more than the delta: both sides rewrote entries from the same tree of 35 independently \u2014 the trunk six (recorded above), this branch two (deferred-finding-decision-reopen, spent-inbox-artifacts-await-deletion, rewritten because road-to-inbox-harvest-residuals was being touched anyway). The sets are disjoint, so the merged tree measures 27 and the gate itself reported the slack. Tightened rather than banked: 29 against a 27-entry tree would absorb the next two undecidable blockers without anything noticing, which is the lag this note has now corrected twice. Lowered 27 \u2192 26 on 2026-08-16 by rewriting `picktier-wire-or-archive` (road-to-inbox-harvest-2026-08-d-top-band-model-economy), and this one is the usual shape the note describes rather than the dedicated-drain shape: the roadmap was being worked for its own Step 4.1 and the entry was rewritten because it was open on the desk. Worth recording for what writing the two fields SURFACED, which is the argument for the requirement: `Recommendation:` forced the question 'which option, and why' and the honest answer turned out not to be one of the three options at all \u2014 the entry defers to a converged council verdict, and the council cannot reach quorum (re-measured 0 of 2 the same day), so the actionable next step is a one-line fix to a user-global config file outside this repository. Five fields had described that situation accurately for two weeks without anyone noticing that the named blocker was infrastructure rather than judgement. Lowered 26 \u2192 22 on 2026-08-18 by road-to-estate-drawdown Step 0.1, which needed the two fields for a different reason and therefore drained six entries as a by-product: the consolidated decision sheet renders a DEFAULT plus that default's provenance per user-owned decision, and 7 of the 21 live entries recorded no `Recommendation:` to derive one from. Six were rewritten to the new standard (skill-activation-window, b-convergence-machine, maintainer-blind-ratings, real-orchestration-usage, human-gated-live-trigger-eval, manual-rubric-rater), each marked `agent-drafted` in the field itself so the sheet can label the provenance rather than presenting a drafted default beside a maintainer-recorded one as the same kind of statement. The seventh, `road-to-gated-reach-followup`'s legacy `> Blocked until \u2026` note, is NOT counted here and was not rewritten: it is a blocked-until note rather than a `### blocker:` entry, so it has no field to carry, and converting it is a roadmap restructure this change did not take. **The tree was RED at 28 against this baseline of 26 before the drain, and again nobody knew** \u2014 the same failure the 34\u219229 paragraph above records, one gate cycle later: two entries had been added on the trunk without the fields, and this gate runs in `task ci`, which no workflow invokes. That is the second recorded instance of this ratchet being red on main undetected, which is an argument about where it runs rather than about its number. `human-gated-live-trigger-eval` also gained a command in `What to do:` (`rule_trigger_eval`, verified to exist at src/scripts/) because the field carried a path-less instruction; the other five needed only the two decidability fields. Lowered 22 -> 20 on 2026-08-19 by road-to-estate-drawdown Phase 2 triage batch 1, and RE-MEASURED after two merges of origin/main rather than carried over: the trunk gained roadmaps in between (the long-horizon pair and the council-fallback set), and the live count over the merged tree is still 20 across 39 roadmaps. Six roadmaps moved to later/, and two of them carried an undecidable open blocker each. The gain is a side effect of parking rather than of drafting a recommendation, so it is recorded as such \u2014 the two entries are unchanged and still undecidable where they sit; they are simply outside the active tree this ratchet scans. Walked down because the gate reported itself loose, and a baseline above the measurement is headroom a later change can spend with nothing objecting. Lowered 20 -> 17 on 2026-08-20 by the road-to-frontend-skill-application closeout, and the shape is the usual one this note describes: the roadmap was being closed for its own reasons and its three open blockers were on the desk anyway. Two (ui-corpus-has-no-ui, enforcement-evidence) left the population by being RESOLVED rather than rewritten, which is the cleanest way to lower this number and the only one that does not add prose; the third (ui-session-capture-window) was rewritten to the new standard. Worth recording what writing the two fields surfaced, since that is the standing argument for the requirement: on ui-corpus-has-no-ui, `Recommendation:` forced the question \"which option, and why\" and the honest answer was that the entry's own premise was false \u2014 it asserted the question could not be answered without a human naming a consumer store, and the stores were already on the machine, readable by the command the entry itself printed. Five fields had described that situation accurately for a week without anyone noticing the blocker was a measurement nobody had run. Tightened rather than banked because 20 against a 17-entry tree is headroom the next three undecidable blockers would spend with nothing objecting, which is the lag the paragraphs above have now corrected three times. Walked 17 -> 12 at the drain/road-to-frontend-skill-application merge, 2026-08-21, on the gate's own loose-ratchet report. Re-measured against the merged tree rather than attributed: sibling drain PRs made entries decidable in parallel, so this branch's share is not separable here and is not claimed. Banked at the measurement rather than left as headroom a later change could spend.\n\nWALKED 12 -> 1 on 2026-08-22 by resolving the two `road-to-evidence-based-adr-governance` blockers (`owner-autonomy-batch`, `authority-coupling-decision`), which were undecidable in exactly the way this ratchet names: each handed a Hard-Floor question to the reader without a landed outcome. Both now carry one. The single remaining violation is `road-to-drain-commands:395` (`merge-authority`), which PR #1517 resolves and archives \u2014 so the honest count for THIS tree is 1 and it is expected to reach 0 once that lands. It is not lowered to 0 here, because a baseline must describe the tree it ships with rather than the tree someone expects next. Lowered 1 -> 0 on 2026-08-22, and the prediction the previous entry made is what closed it: the single remaining violation was `road-to-drain-commands:395` (`merge-authority`), and PR #1517 archived that roadmap, taking the entry out of the scanned population. Nothing was rewritten to reach zero -- the record left the active tree with its blocker RESOLVED, which is the cleanest of the two ways this number goes down and the only one that adds no prose. Zero is now the ceiling: any new open blocker missing `Recommendation:`, `If you do nothing:` or an actionable `What to do:` reds this gate immediately, with no headroom at all.",
+      "falsifier": "./scripts-run src/scripts/lint_roadmap_blockers --quiet",
+      "retired_on": "2026-08-22",
+      "retired_because": "Walked to 0 and therefore removed from `gates`: an entry recording no debt is what gate_baseline.test.ts refuses, and a missing entry is not weaker \u2014 evaluateGate treats an absent baseline as ok only when the live count is 0, which is the same ceiling the note below asked for. Kept here because the note is dated history, not commentary; nothing reads this key at runtime."
+    }
+  }
 }

package/src/config/gitignore-block.txt

@@ -2,8 +2,21 @@
 # install). After the Phase-5 migration the entire `.augment/`, `.claude/`,
 # `.cursor/` trees live under `~/.event4u/agent-config/`; consumers carry
 # only `agents/overrides/`. Pre-Phase-5 entries kept as a defence-in-depth
-# ignore list for legacy installs that have not migrated yet — Phase 5
-# unstages them on first migrate run.
+# ignore list for legacy installs that have not migrated yet.
+#
+# CORRECTED 2026-08-21 (road-to-merge-hotspot-drawdown 2.2): this header used to
+# end "— Phase 5 unstages them on first migrate run", and NOTHING DOES. Verified
+# on the live tree: `cmd_migrate.ts` refreshes this block
+# (`_update_gitignore`, :619-647) and performs no git index operation of any
+# kind, and no executable path anywhere in the repo runs `git rm --cached` —
+# `check_tracked_but_ignored.ts:80` and `check_no_local_settings_committed.ts:85`
+# only PRINT the command, and `/sync-gitignore:fix` states in as many words that
+# git-ops are user-owned and must not be run automatically. So an entry added
+# here ignores a path for the future and leaves an already-committed file
+# tracked; a consumer who has committed it must unstage it themselves. The false
+# claim mattered because a consumer-rollout plan was built on it as precedent:
+# an auto-unstage would be a new mechanism AND a deliberate reversal of the
+# user-owned-git-ops design, not an extension of something already shipped.
 .augment/skills/
 .augment/commands/
 .augment/guidelines/

package/src/config/hook-latency-budget.json

@@ -49,6 +49,11 @@
         "tool_calls": 10,
         "observe_only": true,
         "p50_ci": null,
+        "observe_only_decision": "b-per-turn-composite-bar option (b) — council 2026-08-20, 2/2 quorum, both seats convergent: \"mark (pre + post) x 10 + ups + stop observe-only for one release, then derive an absolute ceiling and a pathology net from that distribution.\" Option (a), adopting the source draft's p50 <= 1.5 s, was refused because no run in this tree produced that number and an invented bar on a summed metric is the flappiest gate available. Option (c), declining the composite, was refused because it leaves D-1 permanently unmeasurable, which is the defect itself.",
+        "observe_only_since_version": "14.6.0",
+        "arming_procedure": "1. Collect the printed composite from every CI gate run across one release cycle — the bench prints it on every run, gated or not. 2. Set p50_ci to an ABSOLUTE cap derived from that distribution, never a creep window: shared CI runners flap, which is why regression_gate.max_regression_pct is already 200. 3. Keep the pathology net separate from the cap, so one spike is caught as a pathology rather than by lowering the cap. 4. Set observe_only to false — no code change is needed to arm it.",
+        "arming_precondition": "At least 10 CI gate readings of the composite, from at least 2 distinct runner sessions. Below that the estimator cannot separate code cost from runner load: the transport-isolation cell added for b-payload-read-parse-dominates read 44-157% at n=12 and 69-74% at n=50 on ONE machine, and a bar set from single-digit readings would carry that same instability.",
+        "_arming_precondition_honesty": "The 10-readings / 2-sessions floor is a STATED MINIMUM, not a derived one — it is the point below which the n=12 instability above was directly observed on a sibling metric in this same roadmap. Revisit-if: a release cycle produces the readings and their spread demands a wider floor, or the composite turns out to be dominated by one slot, in which case that slot's own cap is the better gate and this row stays observe-only permanently with that recorded.",
         "_observe_only_note": "While observe_only is true the bench PRINTS the composite on every run, including the CI gate run, and never fails on it. Setting p50_ci to a number and observe_only to false is the whole of step 4.2 — no code change is needed to arm it. Do not arm it from a single run: the composite sums four slot measurements and multiplies two of them by tool_calls, so it accumulates runner variance faster than any single slot (risk 5), which is why the same absolute-cap-plus-pathology-net posture the slots use applies here too."
     },
     "regression_gate": {