gaia-framework 1.57.0 → 1.57.2

package/_gaia/lifecycle/workflows/4-implementation/val-refresh-ground-truth/checklist.md

@@ -1,48 +0,0 @@
----
-title: 'Val Refresh Ground Truth Validation'
-validation-target: 'Ground truth refresh workflow'
----
-## Pre-conditions
-- [ ] Validator agent persona exists at _gaia/lifecycle/agents/validator.md
-- [ ] Slash command registered at .claude/commands/gaia-refresh-ground-truth.md
-- [ ] Ground-truth-management skill available for JIT loading
-
-## Workflow Structure
-- [ ] workflow.yaml created with agent: validator and --incremental parameter
-- [ ] instructions.xml implements complete scan protocol
-- [ ] checklist.md validates workflow completeness
-
-## Scan Protocol
-- [ ] All 11 inventory targets scanned
-- [ ] Section-by-section progress shown during scan
-- [ ] Metadata extracted from each scanned file
-
-## Ground Truth Output
-- [ ] ground-truth.md updated with verified counts and locations
-- [ ] Last-refresh timestamp in header
-- [ ] Entries organized by category
-- [ ] REMOVED status for deleted files (not silently purged)
-
-## Diff Report
-- [ ] Added/removed/updated counts generated
-- [ ] Total entry count included
-- [ ] Report presented to user
-- [ ] Report logged to decision-log.md with date
-
-## Modes
-- [ ] Full refresh scans all targets (default)
-- [ ] Incremental refresh filters by last-refresh timestamp
-- [ ] Incremental limitation documented (misses deletions)
-
-## First-Run
-- [ ] Creates _memory/validator-sidecar/ if missing
-- [ ] Initializes ground-truth.md, decision-log.md, conversation-context.md
-
-## Skill Loading
-- [ ] ground-truth-management sections loaded JIT
-- [ ] Sections: full-refresh, incremental-refresh, entry-structure, conflict-resolution, token-budget
-
-## Integration
-- [ ] Manifest entry exists in workflow-manifest.csv
-- [ ] Works identically standalone or as sub-step
-- [ ] No special-case logic for invocation context

package/_gaia/lifecycle/workflows/4-implementation/val-refresh-ground-truth/instructions.xml

@@ -1,125 +0,0 @@
-<workflow name="val-refresh-ground-truth">
-<critical>
-  <mandate>Ground truth accuracy is foundational — every other Val workflow depends on it</mandate>
-  <mandate>Never silently delete entries — mark removed files with REMOVED status</mandate>
-  <mandate>Load ground-truth-management skill sections JIT — never pre-load</mandate>
-  <mandate>Show section-by-section progress during scanning</mandate>
-  <mandate>Behavior must be identical whether called standalone or as sub-step from another workflow</mandate>
-</critical>
-
-<step n="1" title="Initialize Validator Sidecar">
-  <action>Check if {memory_path}/validator-sidecar/ directory exists</action>
-  <action if="directory missing">Create {memory_path}/validator-sidecar/ directory</action>
-  <action if="ground-truth.md missing">Create {memory_path}/validator-sidecar/ground-truth.md with empty header containing last-refresh timestamp set to "never"</action>
-  <action if="decision-log.md missing">Create {memory_path}/validator-sidecar/decision-log.md with header: "# Val Decision Log" and empty entries</action>
-  <action if="conversation-context.md missing">Create {memory_path}/validator-sidecar/conversation-context.md with header: "# Val Conversation Context" and empty rolling state</action>
-</step>
-
-<step n="2" title="Determine Refresh Mode">
-  <action>Check if --incremental flag was passed</action>
-  <action if="incremental">Set mode to incremental — only scan files modified since last-refresh timestamp from ground-truth.md header</action>
-  <action if="not incremental">Set mode to full — scan all targets completely. Full refresh catches deletions and renames that incremental would miss.</action>
-</step>
-
-<step n="3" title="Load Entry Structure">
-  <action>Load ground-truth-management skill section: entry-structure (JIT)</action>
-  <action>Use entry-structure format for all ground truth entries written in subsequent steps</action>
-</step>
-
-<step n="4" title="Parse Previous State">
-  <action>Read existing {memory_path}/validator-sidecar/ground-truth.md</action>
-  <action>Extract last-refresh timestamp from header</action>
-  <action>Parse all existing entries into a lookup map keyed by file path for diff comparison</action>
-  <action if="incremental mode">Filter scan targets to only files modified after last-refresh timestamp</action>
-</step>
-
-<step n="5" title="Scan Inventory Targets">
-  <action>Load ground-truth-management skill section based on mode: full-refresh or incremental-refresh (JIT)</action>
-  <action>Scan the following 6 inventory targets, showing section-by-section progress to the user after each target completes.</action>
-
-  <action title="Exclusion list">
-    ALWAYS exclude these directories and files from scanning — they are framework internals, not project code:
-    _gaia/, .claude/, bin/, _memory/, node_modules/, .git/, build/, dist/, .DS_Store, *.lock
-  </action>
-
-  <action title="Target 1: Project Source Files">
-    Scan {project-path}/**/* (excluding the exclusion list above)
-    Extract: file inventory, directory structure, languages used, entry points
-    Report progress: "Scanning project source files... found N files across N directories."
-  </action>
-
-  <action title="Target 2: Project Config Files">
-    Scan {project-path}/*.{json,yaml,yml,toml,xml,env.example} (root-level config files)
-    Extract: config keys, settings, dependency declarations
-    Report progress: "Scanning project config files... found N config files."
-  </action>
-
-  <action title="Target 3: Project Package Manifests">
-    Scan {project-path}/**/package.json, pubspec.yaml, pom.xml, build.gradle, requirements.txt, Cargo.toml, go.mod, Gemfile, *.csproj (whichever exist)
-    Extract: dependencies, versions, scripts, build targets
-    Report progress: "Scanning package manifests... found N manifests."
-  </action>
-
-  <action title="Target 4: Planning Artifacts">
-    Scan {project-root}/docs/planning-artifacts/*.md
-    Extract: artifact name, type, date
-    Report progress: "Scanning planning artifacts... found N artifacts."
-  </action>
-
-  <action title="Target 5: Implementation Artifacts">
-    Scan {project-root}/docs/implementation-artifacts/*.md
-    Extract: artifact name, type, story key if applicable
-    Report progress: "Scanning implementation artifacts... found N artifacts."
-  </action>
-
-  <action title="Target 6: Test Artifacts">
-    Scan {project-root}/docs/test-artifacts/*.md
-    Extract: artifact name, type, coverage area
-    Report progress: "Scanning test artifacts... found N artifacts."
-  </action>
-</step>
-
-<step n="6" title="Compare and Detect Changes">
-  <action>Compare scan results against previous state from Step 4</action>
-  <action>Classify each entry as: ADDED (new file not in previous state), UPDATED (file exists but metadata changed), UNCHANGED (no changes detected)</action>
-  <action if="full mode">For entries in previous state not found in scan results: mark as REMOVED with detection date (e.g., "REMOVED (file deleted, detected 2026-03-19)"). Do NOT silently delete entries.</action>
-  <action if="incremental mode">Skip deletion detection — incremental mode cannot detect deletions. This is a documented limitation.</action>
-  <action>Load ground-truth-management skill section: conflict-resolution (JIT) if any conflicts are detected between scan results and existing entries</action>
-</step>
-
-<step n="7" title="Write Ground Truth">
-  <action>Update {memory_path}/validator-sidecar/ground-truth.md with all scan results</action>
-  <action>Write header with last-refresh timestamp set to current date/time</action>
-  <action>Organize entries by category: Agents, Workflows, Skills, Commands, Manifests, Config, Artifacts</action>
-  <action>Include verified counts, locations, and structural patterns for each category</action>
-  <action>Preserve REMOVED entries with their detection dates — do not purge</action>
-</step>
-
-<step n="8" title="Generate Diff Report">
-  <action>Generate diff/delta report summarizing changes since last refresh</action>
-  <action>Include counts by category: added, removed, updated entries</action>
-  <action>Include total entry count across all categories</action>
-  <action>Format: "Added: N new workflows, N new agents. Removed: N deleted commands. Updated: N file paths changed. Total entries: N."</action>
-  <action>Present the full diff report to the user</action>
-</step>
-
-<step n="9" title="Log to Decision Log">
-  <action>Append the diff/delta report to {memory_path}/validator-sidecar/decision-log.md</action>
-  <action>Include date, refresh mode (full or incremental), and summary</action>
-  <action>Format: "## Refresh — {date} ({mode})\n{summary}"</action>
-</step>
-
-<step n="10" title="Check Token Budget">
-  <action>Load ground-truth-management skill section: token-budget (JIT)</action>
-  <action>Estimate token count of ground-truth.md</action>
-  <action>If token count exceeds budget threshold: trigger archival of oldest REMOVED entries per token-budget skill guidance</action>
-  <action>Report token usage to user</action>
-</step>
-
-<step n="11" title="Present Results">
-  <template-output file="{memory_path}/validator-sidecar/ground-truth.md">
-    Ground truth refreshed with verified inventory of all framework components.
-    Includes last-refresh timestamp, categorized entries, and REMOVED markers for deleted files.
-  </template-output>
-</step>
-</workflow>

package/_gaia/lifecycle/workflows/4-implementation/val-refresh-ground-truth/workflow.yaml

@@ -1,31 +0,0 @@
-name: val-refresh-ground-truth
-description: 'Rescan filesystem and update verified facts in ground-truth.md'
-module: lifecycle
-agent: validator
-execution_mode: normal
-config_resolved: "{installed_path}/.resolved/val-refresh-ground-truth.yaml"
-config_source: "{project-root}/_gaia/lifecycle/config.yaml"
-installed_path: "{project-root}/_gaia/lifecycle/workflows/4-implementation/val-refresh-ground-truth"
-instructions: "{installed_path}/instructions.xml"
-validation: "{installed_path}/checklist.md"
-parameters:
-  incremental:
-    flag: "--incremental"
-    description: "Only scan files modified since last refresh timestamp. Full refresh is default."
-    default: false
-required_skill_sections:
-  - "ground-truth-management:full-refresh"
-  - "ground-truth-management:incremental-refresh"
-  - "ground-truth-management:entry-structure"
-  - "ground-truth-management:conflict-resolution"
-  - "ground-truth-management:token-budget"
-input_file_patterns:
-  ground_truth:
-    whole: "{memory_path}/validator-sidecar/ground-truth.md"
-    load_strategy: "FULL_LOAD"
-on_error:
-  missing_file: "ask_user"
-  unresolved_variable: "halt"
-
-output:
-  primary: "{memory_path}/validator-sidecar/ground-truth.md"

package/_gaia/lifecycle/workflows/4-implementation/val-validate-artifact/checklist.md

@@ -1,54 +0,0 @@
-# val-validate-artifact — Workflow Checklist
-
-## File Structure
-- [ ] `workflow.yaml` exists in `val-validate-artifact/`
-- [ ] `instructions.xml` exists in `val-validate-artifact/`
-- [ ] `checklist.md` exists in `val-validate-artifact/`
-
-## workflow.yaml Verification
-- [ ] `name: val-validate-artifact`
-- [ ] `module: lifecycle`
-- [ ] `agent: validator`
-- [ ] `model_override: opus` — mandatory, not configurable (ADR-012)
-- [ ] `installed_path` correctly set
-- [ ] `instructions` references `{installed_path}/instructions.xml`
-- [ ] `validation` references `{installed_path}/checklist.md`
-- [ ] `input_file_patterns` includes target artifact and ground-truth paths
-- [ ] `on_error` handles missing ground truth gracefully (continue without cross-reference)
-- [ ] `on_error` halts on missing artifact
-- [ ] `output.primary` references the target artifact itself
-
-## instructions.xml Verification
-- [ ] Exactly 8 steps present
-- [ ] Step 1: Parse Artifact — heading structure, section map, chunking strategy
-- [ ] Step 2: Extract Claims — JIT loads `claim-extraction` section, handles no-claims edge case
-- [ ] Step 3: Filesystem Verify — JIT loads `filesystem-verification` section
-- [ ] Step 4: Cross-Reference Ground Truth — JIT loads `cross-reference` section, handles missing/empty ground truth
-- [ ] Step 5: Classify Findings — JIT loads `severity-classification` section, three severity levels (CRITICAL/WARNING/INFO)
-- [ ] Step 6: Discussion Loop — `template-output` checkpoint, user confirmation enforced
-- [ ] Step 7: Write Approved Findings — JIT loads `findings-formatting` section, replaces existing findings section
-- [ ] Step 8: Save to Val Memory — auto-saves decision-log + conversation-context (no user prompt)
-
-## Skill Section References
-- [ ] `claim-extraction` referenced in Step 2
-- [ ] `filesystem-verification` referenced in Step 3
-- [ ] `cross-reference` referenced in Step 4
-- [ ] `severity-classification` referenced in Step 5
-- [ ] `findings-formatting` referenced in Step 7
-
-## Edge Cases
-- [ ] No-claims artifact: graceful completion with "No factual claims identified" message
-- [ ] Missing ground truth: filesystem verification runs, note included in findings
-- [ ] Empty ground truth: same behavior as missing
-- [ ] All findings disputed: no findings written, graceful completion
-- [ ] Existing `## Validation Findings` section: replaced entirely (not appended)
-
-## Quality Checks
-- [ ] Diplomatic framing enforced (Val proposes, user decides)
-- [ ] User confirmation mandatory before writing findings (template-output in Step 6)
-- [ ] `model_override: opus` enforced — workflow fails clearly if opus unavailable
-- [ ] Findings include timestamp and workflow version
-- [ ] `<next-step>` block references `/gaia-val-validate` (no /gaia-val-save — auto-save is built-in)
-
-## Integration
-- [ ] `workflow-manifest.csv` has `val-validate-artifact` entry with correct fields

package/_gaia/lifecycle/workflows/4-implementation/val-validate-artifact/instructions.xml

@@ -1,147 +0,0 @@
-<workflow name="val-validate-artifact">
-<critical>
-  <mandate>Val NEVER writes findings to an artifact without explicit user confirmation. The discussion loop (Step 6) is a MANDATORY interaction point.</mandate>
-  <mandate>model_override: opus is architecturally mandated (ADR-012). If the current model is not opus, HALT with a clear error — do NOT silently degrade validation quality.</mandate>
-  <mandate>Findings must be framed diplomatically — Val proposes, the user decides. Use neutral language: "appears to be", "may be outdated", "could not verify" rather than accusatory phrasing.</mandate>
-  <mandate>Each skill section is loaded JIT and released before the next step. Never pre-load skill sections.</mandate>
-  <mandate>For large artifacts, validate section-by-section using the heading map from Step 1. Never load the entire artifact at once for validation.</mandate>
-</critical>
-
-<step n="1" title="Parse Artifact">
-  <action>Read the target artifact at {artifact_path}</action>
-  <action>Parse the heading structure (##, ###, ####) into a section map: for each section, record heading level, title, and line range</action>
-  <action>Determine chunking strategy based on artifact size:
-    — Small artifact (under 200 lines): treat as single chunk, validate all at once
-    — Medium artifact (200–600 lines): chunk by top-level sections (## headings)
-    — Large artifact (over 600 lines): chunk by second-level sections (### headings) for finer granularity</action>
-  <action>If the artifact has no headings (e.g., flat content): treat the entire artifact as a single chunk</action>
-  <action>Present the section map to confirm scope: "{N} sections identified, {M} chunks for validation"</action>
-</step>
-
-<step n="2" title="Extract Claims">
-  <!-- SKILL-REF: validation-patterns.md SECTION: claim-extraction -->
-  <action>Load the "claim-extraction" section from the validation-patterns skill (JIT)</action>
-  <action>For each chunk from Step 1, extract all verifiable factual claims:
-    — File paths: any reference to a file or directory (e.g., "_gaia/dev/agents/", "docs/planning-artifacts/prd.md")
-    — Component counts: numerical assertions about how many agents, workflows, skills, etc. exist
-    — Agent/workflow/skill references: named references to framework components
-    — FR/ADR cross-references: requirement IDs (FR-*) and architectural decision references (ADR-*)
-    — Version numbers: framework version, module versions, dependency versions
-    — Structural assertions: claims about directory structure, file contents, or configuration values</action>
-  <action>For each claim, record: claim text, source section, source line (approximate), claim type</action>
-  <action>If no verifiable factual claims are found in the entire artifact: produce a report stating "No factual claims identified for verification" and skip to workflow completion — do NOT write an empty findings section</action>
-  <action>Release the "claim-extraction" skill section from context</action>
-</step>
-
-<step n="3" title="Filesystem Verify">
-  <!-- SKILL-REF: validation-patterns.md SECTION: filesystem-verification -->
-  <action>Load the "filesystem-verification" section from the validation-patterns skill (JIT)</action>
-  <action>For each extracted claim, verify against the actual filesystem:
-    — File path claims: check if the file/directory exists at the stated path using glob or direct read
-    — Count claims: enumerate actual items and compare against the stated count
-    — Content assertions: read the referenced file and verify the stated content matches
-    — Structural claims: verify directory structures match the described layout</action>
-  <action>For each claim, record verification result: VERIFIED (claim matches reality) or FAILED (claim does not match) with evidence</action>
-  <action>Release the "filesystem-verification" skill section from context</action>
-</step>
-
-<step n="4" title="Cross-Reference Ground Truth">
-  <!-- SKILL-REF: validation-patterns.md SECTION: cross-reference -->
-  <action>Check if {memory_path}/validator-sidecar/ground-truth.md exists and is non-empty</action>
-  <action>If ground-truth.md is missing or empty: skip cross-reference checks. Record a note: "Ground truth not available — cross-reference checks skipped." Proceed to Step 5. Filesystem verification results from Step 3 are still valid.</action>
-  <action>If ground-truth.md exists and is non-empty:
-    — Load the "cross-reference" section from the validation-patterns skill (JIT)
-    — For each claim that was VERIFIED in Step 3, cross-reference against ground truth:
-      — Check if ground truth contains a contradicting fact
-      — Check if ground truth has a more recent or more precise version of the same fact
-      — Flag any discrepancies between the artifact claim and ground truth
-    — Release the "cross-reference" skill section from context</action>
-</step>
-
-<step n="5" title="Classify Findings">
-  <!-- SKILL-REF: validation-patterns.md SECTION: severity-classification -->
-  <action>Load the "severity-classification" section from the validation-patterns skill (JIT)</action>
-  <action>Compile all findings from Steps 3 and 4 — only FAILED verifications and ground-truth discrepancies</action>
-  <action>Classify each finding into exactly one severity level:
-    — **CRITICAL**: Wrong file path (file does not exist), incorrect count (stated N but actual is M), broken reference (FR/ADR/component does not exist), contradicts ground truth on a verified fact
-    — **WARNING**: Outdated reference (file exists but content has changed), stale data (version number is behind), ground truth has newer information
-    — **INFO**: Style suggestion (naming inconsistency), minor inconsistency (non-breaking), cosmetic discrepancy</action>
-  <action>If no findings were produced (all claims verified, no discrepancies): report "All {N} claims verified — no findings." Skip Steps 6 and 7 — proceed to workflow completion.</action>
-  <action>Release the "severity-classification" skill section from context</action>
-</step>
-
-<step n="6" title="Discussion Loop">
-  <action>Present all classified findings to the user in a structured table:
-
-  | # | Severity | Section | Claim | Finding | Evidence |
-  |---|----------|---------|-------|---------|----------|
-  (one row per finding)
-
-  Summary: {total} findings — {critical_count} CRITICAL, {warning_count} WARNING, {info_count} INFO</action>
-
-  <template-output file="{memory_path}/validator-sidecar/val-findings-draft.md">
-    Draft findings table with all classified findings, severity counts, and evidence for each.
-    This is a DRAFT — only user-approved findings will be written to the artifact.
-  </template-output>
-
-  <action>For each finding, ask the user to confirm or dispute:
-    — APPROVE: finding is correct, include in written output
-    — DISPUTE: finding is incorrect or not applicable, exclude from written output
-    — EDIT: modify the finding description before approving</action>
-  <action>Only findings explicitly approved by the user proceed to Step 7</action>
-  <action>If the user disputes ALL findings: report "All findings disputed — no changes written to artifact." Complete the workflow gracefully.</action>
-</step>
-
-<step n="7" title="Write Approved Findings">
-  <!-- SKILL-REF: validation-patterns.md SECTION: findings-formatting -->
-  <action>Load the "findings-formatting" section from the validation-patterns skill (JIT)</action>
-  <action>Check if the target artifact already contains a "## Validation Findings" section</action>
-  <action>If an existing "## Validation Findings" section is found: replace it entirely with the new findings. Do NOT append — each validation run is a fresh assessment. Prior findings are preserved in git history.</action>
-  <action>Write the approved findings to the target artifact under a new "## Validation Findings" section:
-
-  ## Validation Findings
-
-  > Validated: {date} | Workflow: val-validate-artifact v{framework_version} | Model: opus
-
-  | # | Severity | Finding | Reference |
-  |---|----------|---------|-----------|
-  (one row per approved finding — severity tag, one-line summary, what was checked)
-
-  Summary: {approved_count} finding(s) from {total_checked} claims verified.
-  </action>
-  <action>Release the "findings-formatting" skill section from context</action>
-</step>
-
-<step n="8" title="Save to Val Memory">
-  <action>Auto-save all validation results to Val's memory sidecar (no user prompt required):
-
-  1. Append to {memory_path}/validator-sidecar/decision-log.md using standardized format:
-     ### [YYYY-MM-DD] Artifact Validation: {artifact_name}
-
-     - **Agent:** validator
-     - **Workflow:** val-validate-artifact
-     - **Sprint:** {sprint_id or "N/A"}
-     - **Status:** recorded
-     - **Related:** {artifact_path}
-
-     Validated artifact: {artifact_path}.
-     Claims checked: {total_claims}. Findings: {total_findings} ({critical_count} critical, {warning_count} warning, {info_count} info).
-     Approved: {approved_count}. Disputed: {disputed_count}.
-     {Brief summary of key findings}
-
-  2. Replace body of {memory_path}/validator-sidecar/conversation-context.md (preserve header above first ---):
-     Last session: Artifact validation of {artifact_name}.
-     Date: {YYYY-MM-DD}. Claims: {total_claims}. Findings: {total_findings}.
-     Approved: {approved_count}. Disputed: {disputed_count}.
-
-  If _memory/validator-sidecar/ directory or files don't exist, create them with standard headers:
-  decision-log.md: "# Val Validator — Decision Log\n\n> Chronological record of validation decisions.\n\n---"
-  conversation-context.md: "# Val Validator — Conversation Context\n\n> Rolling summary of the most recent validation session.\n\n---\n\nNo sessions recorded yet."
-  If writing fails, log warning and continue — memory save is non-blocking.</action>
-</step>
-<next-step>
-  Suggested next steps:
-  - /gaia-val-validate — run validation again on a different artifact
-</next-step>
-
-</workflow>

package/_gaia/lifecycle/workflows/4-implementation/val-validate-artifact/test-structure.sh

@@ -1,116 +0,0 @@
-#!/bin/bash
-# Structural validation tests for E8-S6: val-validate-artifact workflow
-# These tests verify all acceptance criteria are met.
-
-WORKFLOW_DIR="_gaia/lifecycle/workflows/4-implementation/val-validate-artifact"
-MANIFEST="_gaia/_config/workflow-manifest.csv"
-ROOT=$(cd "$(dirname "$0")/../../../../.." && pwd)
-PASS=0
-FAIL=0
-
-pass() { echo "  ✓ $1"; PASS=$((PASS + 1)); }
-fail() { echo "  ✗ $1"; FAIL=$((FAIL + 1)); }
-
-cd "$ROOT"
-
-echo "=== E8-S6 Structural Tests ==="
-echo ""
-
-# AC1: Directory contains three files
-echo "AC1: Workflow directory structure"
-[ -f "$WORKFLOW_DIR/workflow.yaml" ] && pass "workflow.yaml exists" || fail "workflow.yaml missing"
-[ -f "$WORKFLOW_DIR/instructions.xml" ] && pass "instructions.xml exists" || fail "instructions.xml missing"
-[ -f "$WORKFLOW_DIR/checklist.md" ] && pass "checklist.md exists" || fail "checklist.md missing"
-
-# AC2: workflow.yaml fields
-echo ""
-echo "AC2: workflow.yaml required fields"
-if [ -f "$WORKFLOW_DIR/workflow.yaml" ]; then
-  grep -q "^name: val-validate-artifact" "$WORKFLOW_DIR/workflow.yaml" && pass "name field correct" || fail "name field incorrect"
-  grep -q "^module: lifecycle" "$WORKFLOW_DIR/workflow.yaml" && pass "module field correct" || fail "module field incorrect"
-  grep -q "^agent: validator" "$WORKFLOW_DIR/workflow.yaml" && pass "agent field correct" || fail "agent field incorrect"
-  grep -q "model_override: opus" "$WORKFLOW_DIR/workflow.yaml" && pass "model_override: opus present" || fail "model_override: opus missing (AC15)"
-  grep -q "installed_path:" "$WORKFLOW_DIR/workflow.yaml" && pass "installed_path present" || fail "installed_path missing"
-  grep -q "instructions:" "$WORKFLOW_DIR/workflow.yaml" && pass "instructions present" || fail "instructions missing"
-  grep -q "validation:" "$WORKFLOW_DIR/workflow.yaml" && pass "validation present" || fail "validation missing"
-  grep -q "input_file_patterns:" "$WORKFLOW_DIR/workflow.yaml" && pass "input_file_patterns present" || fail "input_file_patterns missing"
-else
-  fail "workflow.yaml not found — skipping field checks"
-fi
-
-# AC3: instructions.xml has exactly 7 steps
-echo ""
-echo "AC3: instructions.xml step structure"
-if [ -f "$WORKFLOW_DIR/instructions.xml" ]; then
-  STEP_COUNT=$(grep -c '<step n=' "$WORKFLOW_DIR/instructions.xml")
-  [ "$STEP_COUNT" -eq 7 ] && pass "Exactly 7 steps found" || fail "Expected 7 steps, found $STEP_COUNT"
-
-  # Verify step titles in order
-  grep -q 'n="1".*Parse Artifact' "$WORKFLOW_DIR/instructions.xml" && pass "Step 1: Parse Artifact" || fail "Step 1 title mismatch"
-  grep -q 'n="2".*Extract Claims' "$WORKFLOW_DIR/instructions.xml" && pass "Step 2: Extract Claims" || fail "Step 2 title mismatch"
-  grep -q 'n="3".*Filesystem Verify' "$WORKFLOW_DIR/instructions.xml" && pass "Step 3: Filesystem Verify" || fail "Step 3 title mismatch"
-  grep -q 'n="4".*Cross-Reference Ground Truth' "$WORKFLOW_DIR/instructions.xml" && pass "Step 4: Cross-Reference Ground Truth" || fail "Step 4 title mismatch"
-  grep -q 'n="5".*Classify Findings' "$WORKFLOW_DIR/instructions.xml" && pass "Step 5: Classify Findings" || fail "Step 5 title mismatch"
-  grep -q 'n="6".*Discussion Loop' "$WORKFLOW_DIR/instructions.xml" && pass "Step 6: Discussion Loop" || fail "Step 6 title mismatch"
-  grep -q 'n="7".*Write Approved Findings' "$WORKFLOW_DIR/instructions.xml" && pass "Step 7: Write Approved Findings" || fail "Step 7 title mismatch"
-else
-  fail "instructions.xml not found — skipping step checks"
-fi
-
-# AC5-9: Skill section references (must match validation-patterns.md section names exactly)
-echo ""
-echo "AC5-9: Skill section references"
-if [ -f "$WORKFLOW_DIR/instructions.xml" ]; then
-  grep -q "claim-extraction" "$WORKFLOW_DIR/instructions.xml" && pass "claim-extraction section referenced" || fail "claim-extraction section NOT referenced"
-  grep -q "filesystem-verification" "$WORKFLOW_DIR/instructions.xml" && pass "filesystem-verification section referenced" || fail "filesystem-verification section NOT referenced"
-  grep -q "cross-reference" "$WORKFLOW_DIR/instructions.xml" && pass "cross-reference section referenced" || fail "cross-reference section NOT referenced"
-  grep -q "severity-classification" "$WORKFLOW_DIR/instructions.xml" && pass "severity-classification section referenced" || fail "severity-classification section NOT referenced"
-  grep -q "findings-formatting" "$WORKFLOW_DIR/instructions.xml" && pass "findings-formatting section referenced" || fail "findings-formatting section NOT referenced"
-else
-  fail "instructions.xml not found — skipping skill section checks"
-fi
-
-# AC8: Ground truth missing/empty handling
-echo ""
-echo "AC8: Ground truth graceful handling"
-if [ -f "$WORKFLOW_DIR/instructions.xml" ]; then
-  grep -q "Ground truth not available" "$WORKFLOW_DIR/instructions.xml" && pass "Missing ground truth message present" || fail "Missing ground truth message NOT found"
-else
-  fail "instructions.xml not found"
-fi
-
-# AC10: No-claims edge case
-echo ""
-echo "AC10: No-claims edge case"
-if [ -f "$WORKFLOW_DIR/instructions.xml" ]; then
-  grep -q "No factual claims" "$WORKFLOW_DIR/instructions.xml" && pass "No-claims message present" || fail "No-claims message NOT found"
-else
-  fail "instructions.xml not found"
-fi
-
-# AC11: Discussion loop has template-output
-echo ""
-echo "AC11: Discussion loop template-output"
-if [ -f "$WORKFLOW_DIR/instructions.xml" ]; then
-  grep -q "template-output" "$WORKFLOW_DIR/instructions.xml" && pass "template-output tag present" || fail "template-output tag NOT found"
-else
-  fail "instructions.xml not found"
-fi
-
-# AC14: workflow-manifest.csv entry
-echo ""
-echo "AC14: workflow-manifest.csv entry"
-if [ -f "$MANIFEST" ]; then
-  grep -q "val-validate-artifact" "$MANIFEST" && pass "val-validate-artifact in manifest" || fail "val-validate-artifact NOT in manifest"
-else
-  fail "workflow-manifest.csv not found"
-fi
-
-# AC15: model_override opus
-echo ""
-echo "AC15: model_override opus (covered in AC2)"
-
-# Summary
-echo ""
-echo "=== Results: $PASS passed, $FAIL failed ==="
-exit $FAIL

package/_gaia/lifecycle/workflows/4-implementation/val-validate-artifact/workflow.yaml

@@ -1,29 +0,0 @@
-name: val-validate-artifact
-description: 'Validate artifact against codebase and ground truth — parse, extract claims, verify, classify, discuss, write findings'
-module: lifecycle
-agent: validator
-model_override: opus
-config_resolved: "{installed_path}/.resolved/val-validate-artifact.yaml"
-config_source: "{project-root}/_gaia/lifecycle/config.yaml"
-installed_path: "{project-root}/_gaia/lifecycle/workflows/4-implementation/val-validate-artifact"
-instructions: "{installed_path}/instructions.xml"
-validation: "{installed_path}/checklist.md"
-input_file_patterns:
-  artifact:
-    whole: "{artifact_path}"
-    load_strategy: "INDEX_GUIDED"
-  ground_truth:
-    whole: "{memory_path}/validator-sidecar/ground-truth.md"
-    load_strategy: "FULL_LOAD"
-quality_gates:
-  pre_start:
-    - check: "artifact_path is provided and file exists"
-      on_fail: "HALT: Target artifact path is required. Usage: /gaia-val-validate <path-to-artifact>"
-    - check: "model_override == opus"
-      on_fail: "HALT: Val requires opus model per ADR-012. Set model_override to opus in /gaia-val-validate workflow.yaml."
-on_error:
-  missing_file: "skip_and_note"
-  unresolved_variable: "halt"
-
-output:
-  primary: "{artifact_path}"

package/_gaia/lifecycle/workflows/4-implementation/val-validate-plan/checklist.md

@@ -1,34 +0,0 @@
----
-title: val-validate-plan Validation Checklist
-validation-target: val-validate-plan workflow
----
-
-# val-validate-plan — Validation Checklist
-
-## Structure
-- [ ] workflow.yaml exists with correct fields (name, module, agent, model_override, instructions, validation)
-- [ ] instructions.xml exists with numbered steps in correct order
-- [ ] checklist.md exists with validation items
-- [ ] model_override set to opus in workflow.yaml
-
-## Protocol Completeness
-- [ ] Step 1: Plan parsing extracts file targets, version bumps, and ADR references
-- [ ] Step 2: File target verification classifies by action verb (Create vs Modify vs Delete)
-- [ ] Step 3: Version bump verification compares plan vs codebase semver values
-- [ ] Step 4: Completeness verification checks for related manifest/config/command files
-- [ ] Step 5: Architecture cross-reference checks ADR alignment (with ground truth fallback)
-- [ ] Step 6: Findings classified by severity (CRITICAL/WARNING/INFO) with discussion loop
-- [ ] Step 7: Approved findings written to plan artifact under Plan Validation Findings section
-- [ ] Step 8: Save to Val Memory — auto-saves decision-log + conversation-context (no user prompt)
-
-## Quality
-- [ ] Val is read-only on plan content (append-only for findings section)
-- [ ] Findings use concise one-line-per-finding table format
-- [ ] Discussion loop allows user to approve, dismiss, or edit individual findings
-- [ ] Ground truth fallback produces INFO finding when not seeded (not a hard error)
-- [ ] Skill sections loaded JIT (claim-extraction, filesystem-verification, severity-classification, findings-formatting)
-
-## Integration
-- [ ] workflow-manifest.csv contains val-validate-plan entry
-- [ ] Slash command file exists at .claude/commands/gaia-val-validate-plan.md
-- [ ] Agent field set to validator