Codex compatibility note:
- Invoke repository skills with
$skill-name in Codex; this mirrored copy rewrites legacy Claude /skill-name references.
- Task tracker mandate: BEFORE executing any workflow or skill step, create/update task tracking for all steps and keep it synchronized as progress changes.
- User-question prompts mean to ask the user directly in Codex.
- Ignore Claude-specific mode-switch instructions when they appear.
- Strict execution contract: when a user explicitly invokes a skill, execute that skill protocol as written.
- Subagent authorization: when a skill is user-invoked or AI-detected and its protocol requires subagents, that skill activation authorizes use of the required
spawn_agent subagent(s) for that task.
- Do not skip, reorder, or merge protocol steps unless the user explicitly approves the deviation first.
- For workflow skills, execute each listed child-skill step explicitly and report step-by-step evidence.
- If a required step/tool cannot run in this environment, stop and ask the user before adapting.
Codex Project-Reference Loading (No Hooks)
Codex uses static project-reference loading instead of runtime-injected project docs.
When coding, planning, debugging, testing, or reviewing, open project docs explicitly using this routing.
Always read:
docs/project-config.json (project-specific paths, commands, modules, and workflow/test settings)
docs/project-reference/docs-index-reference.md (routes to the full docs/project-reference/* catalog)
docs/project-reference/lessons.md (always-on guardrails and anti-patterns)
Missing/stale context route: If docs/project-config.json, the docs index, lessons.md, CLAUDE.md, AGENTS.md, or any task-required reference doc is missing or stale, auto-run $project-init or the narrow setup route ($project-config, $docs-init, $scan-all, $scan --target=<key>, $claude-md-init) before ordinary project-specific work. If Codex mirrors or AGENTS.md are missing/stale, ask the user to run $sync-codex; do not auto-run it.
Situation-based docs:
- Project structure/architecture/tech-stack/deployment/setup (any layer — backend, frontend, or infra):
project-structure-reference.md
- Backend/CQRS/API/domain/entity changes:
backend-patterns-reference.md, domain-entities-reference.md
- Frontend/UI/styling/design-system:
frontend-patterns-reference.md, scss-styling-guide.md, design-system/README.md
- Spec authoring,
docs/specs/ pathing, or TC format: feature-spec-reference.md, spec-system-reference.md, spec-principles.md
- Behavior/public-contract changes or spec-test-code sync:
workflow-spec-test-code-cycle-reference.md plus the spec docs above
- Derived spec indexes/ERDs/reimplementation guides:
spec-system-reference.md and source Feature Specs under docs/specs/
- Integration test implementation/review:
integration-test-reference.md
- E2E test implementation/review:
e2e-test-reference.md
- Code review/audit work:
code-review-rules.md plus domain docs above based on changed files
Do not read all docs blindly. Start from docs-index-reference.md, then open only relevant files for the task.
Quick Summary
Goal: Generate a detailed, AI-analyzed release document from git history — by time range (e.g., last 30 days) or custom comparison (e.g., v1.0.0..HEAD) or a user-described focus area.
Workflow:
- Resolve Range — Determine the git range from user input
- Dump Artifacts — Save log, file-status, and full diff to
docs/release/ BEFORE analyzing
- Categorize Changes — Group changed files by framework area (hooks, skills, agents, workflows, principles, config, docs)
- Read Key Diffs — For each category, read the most significant changes via
git show or git diff
- Generate Doc — Write structured markdown release doc to
docs/release/
- Validate — Verify completeness and accuracy against the artifact files
Key Rules:
- Dump first, read second — NEVER analyze a diff you haven't saved to a file first
- Group by theme — Do not list commits chronologically; group by what area they impact
- Cite evidence — Every section claim must reference a specific commit hash or file path
- Custom focus honored — If user provides a focus prompt, weight that area more deeply
- Draft status — All generated docs are Draft; add "Status: Draft" to the header
Release Doc Skill
Generate a detailed, narrative release document by analyzing git history over a time range or between two refs.
Invocation
$release-doc [--days N] [--since DATE] [--range base..head] [--focus "custom prompt"] [--output path]
Examples:
# Last 30 days
$release-doc --days 30
# Since a specific date
$release-doc --since 2026-03-15
# Between two tags or commits
$release-doc --range v1.0.0..HEAD
# With custom focus area
$release-doc --days 30 --focus "what changed in hooks and workflow enforcement"
# Specify output file
$release-doc --days 30 --output docs/release/release-2026-Q1.md
Differences from $release-notes and $changelog
| Feature |
$release-doc |
$release-notes |
$changelog |
| Time-based ranges |
✅ --days N / --since |
❌ ref-to-ref only |
❌ PR/commit only |
| AI thematic analysis |
✅ Groups by framework area |
❌ Commit-type categories |
❌ File-by-file |
| Dumps artifacts to files |
✅ Mandatory before analysis |
❌ Pipes in memory |
❌ In memory |
| Custom focus prompt |
✅ --focus "..." |
❌ |
❌ |
| Narrative sections |
✅ Context + rationale |
❌ Bullet list only |
❌ Bullet list only |
| Framework-level docs |
✅ hooks/skills/agents |
❌ App features |
❌ App features |
Step 1: Resolve Range
Determine the git range from the user's input:
# Time-based (--days N or --since DATE)
SINCE_DATE=$(date -d "-30 days" +%Y-%m-%d) # Linux
SINCE_DATE=$(date -v-30d +%Y-%m-%d) # macOS
git log --since="$SINCE_DATE" --oneline --format="%H %ad %s" --date=short
# Ref-based (--range base..head)
git log v1.0.0..HEAD --oneline --format="%H %ad %s" --date=short
# Count commits in range
git log --since="$SINCE_DATE" --oneline | wc -l
Identify the boundary commits:
# Oldest commit in range (for diff base)
OLDEST=$(git log --since="$SINCE_DATE" --format="%H" | tail -1)
# Newest
NEWEST=HEAD
Step 2: Dump Git Artifacts (MANDATORY BEFORE ANALYSIS)
CRITICAL: Run ALL dumps before reading ANY content. Use background runs for large diffs.
# Create output directory
mkdir -p docs/release
# 1. Full git log with bodies
git log --since="$SINCE_DATE" --format="%H %ad %s%n%b" --date=short \
> docs/release/git-log-{PERIOD}.txt
# 2. File-level change status (A/M/D)
git diff ${OLDEST}^..HEAD --name-status \
> docs/release/diff-file-status-{PERIOD}.txt
# 3. Diff stat summary (line counts)
git diff ${OLDEST}^..HEAD --stat \
> docs/release/diff-stat-{PERIOD}.txt
# 4. Full consolidated diff (background — may be large)
git diff ${OLDEST}^..HEAD \
> docs/release/git-diff-{PERIOD}-full.txt &
echo "Artifacts saved. Commit count: $(cat docs/release/git-log-{PERIOD}.txt | grep -c '^[a-f0-9]\{40\}')"
Replace {PERIOD} with a human-readable period string, e.g., 30d, 2026-03-15-to-2026-04-14.
Verify all artifact files exist and are non-empty before proceeding.
Step 3: Categorize Changes by Framework Area
Read docs/release/diff-file-status-{PERIOD}.txt and group files into categories:
# Count by area
grep -c "^[AMD]\s*.claude/hooks/" docs/release/diff-file-status-{PERIOD}.txt
grep -c "^[AMD]\s*.claude/skills/" docs/release/diff-file-status-{PERIOD}.txt
grep -c "^[AMD]\s*.claude/agents/" docs/release/diff-file-status-{PERIOD}.txt
grep -c "^[AMD]\s*.claude/workflows" docs/release/diff-file-status-{PERIOD}.txt
grep -c "^[AMD]\s*.claude/docs/" docs/release/diff-file-status-{PERIOD}.txt
grep -c "^[AMD]\s*docs/" docs/release/diff-file-status-{PERIOD}.txt
grep -c "^[AMD]\s*CLAUDE.md" docs/release/diff-file-status-{PERIOD}.txt
Standard category map for easy-claude projects:
| File Path Pattern |
Category |
.claude/hooks/** |
Hook Enhancements |
.claude/hooks/lib/** |
Hook Library |
.claude/skills/** |
Skills |
.claude/agents/** |
Agent Definitions |
.claude/workflows/** |
Workflow Orchestration |
.claude/workflows.json |
Workflow Registry |
.claude/docs/** |
Framework Documentation |
.claude/scripts/** |
Tooling & Scripts |
docs/project-reference/** |
Project Reference Docs |
CLAUDE.md |
Principles & Core Rules |
.claude/.ck.json / settings.json |
Configuration |
For non-easy-claude projects, derive categories from the file path patterns actually present.
Step 4: Analyze Key Changes Per Category
For each category with significant changes (>5 files or >200 lines), read representative diffs:
# Most impactful commits (by files changed)
git log --since="$SINCE_DATE" --format="%H %s" --name-only | \
grep -A100 "^[a-f0-9]\{40\}" | head -200
# Show specific commit detail
git show {COMMIT_HASH} --stat --format="%s%n%b"
# Diff for a specific file across the range
git diff ${OLDEST}^..HEAD -- .claude/hooks/init-prompt-gate.cjs
Analysis questions for each category:
- Hooks: What new gates or enforcement were added? What context injection changed? What was the trigger behavior before vs after?
- Skills: What new skills were created? What protocols were updated in existing skills? What was removed or merged?
- Agents: What agent definitions changed? Did tool lists or dispatch rules change?
- Workflows: What workflow steps were added/removed? Did sequence order change? Were any workflows merged or deleted?
- Principles: What rules in CLAUDE.md changed? Were confidence thresholds tightened? New hard gates added?
- Config: What new schema fields? What validation was added?
- Scripts/Tooling: What new utilities? What improved or automated?
Evidence requirement: For every claim in the doc, record {commit_hash}:{file_path} as your source.
Step 5: Handle Custom Focus Prompt
If user provided --focus "...", apply additional depth to that area:
- Grep all changed files matching the focus keywords
- Read the full diff for matching files (not just stat)
- Include a dedicated top-level section for the focus area in the output doc
- Cross-reference related changes in other areas (e.g., a new skill + its hook + its workflow entry)
Step 6: Generate the Release Doc
Write to docs/release/release-notes-{PERIOD}.md (or --output path).
Required structure:
# Release Notes — {Project Name}
**Period:** {START} → {END}
**Commits:** {N} commits | **Files changed:** {N} | **Insertions:** +{N} | **Deletions:** −{N}
**Status:** Draft
---
## Executive Summary
3-4 sentences. The most transformative changes, their business/developer impact, and the overall trajectory.
---
## {Category 1}: {Theme Name}
### What Changed
{Narrative description — not bullet-list of commits}
### Why It Matters
{Impact on how developers/AI uses the system}
### Key Additions / Modifications
| Item | Type | Effect |
| ---- | ---- | ------ |
---
## {Category 2}: ...
[repeat for each significant category]
---
## New Skills Added
| Skill | Command | Purpose |
| ----- | ------- | ------- |
## Skills Removed or Merged
| Skill | Reason |
| ----- | ------ |
## Principles & Rules Updated
| Rule | Before | After | Reason |
| ---- | ------ | ----- | ------ |
---
## Summary Statistics
| Category | Count |
| -------------- | ----- |
| Commits | {N} |
| Files changed | {N} |
| New skills | {N} |
| Skills removed | {N} |
| New hooks | {N} |
| Hooks modified | {N} |
| Agents updated | {N} |
| Net lines | +{N} |
---
_Generated: {DATE} | Branch: {BRANCH} | Range: {OLDEST_HASH}..{HEAD_HASH}_
_Artifacts: `docs/release/git-log-{PERIOD}.txt`, `docs/release/diff-file-status-{PERIOD}.txt`_
Step 7: Validate
After writing, verify:
Configuration: Category Map Override
For projects with custom directory structures, override the category map via project-config.json:
{
"releaseDoc": {
"categoryMap": {
"{api-source-root}/**": "API Layer",
"{ui-source-root}/**": "Frontend",
"migrations/**": "Database Schema"
},
"outputDir": "docs/release",
"focusAreas": ["hooks", "skills", "agents"]
}
}
Integration with Other Skills
$release-notes — Use after $release-doc to generate a consumer-facing version (tag-to-tag, conventional commits)
$changelog — Use for individual feature changelog entries; $release-doc is for multi-week summaries
$graph-blast-radius — Run before release to assess impact of the most changed files
$docs-update — Update project reference docs after analyzing what changed
$commit — Commit the generated release doc
Troubleshooting
Diff too large for context
Solution: The skill requires artifact files be written first — never read the full diff into context. Use git show {hash} for individual commits and grep -l to narrow file scope.
Time range yields 0 commits
# Verify date format (ISO 8601)
git log --since="2026-03-15" --oneline | head -5
# If empty, check git's date interpretation:
git log --format="%ad" --date=short | head -5
Non-easy-claude project structure
The skill auto-derives categories from file paths. For repositories without .claude/, group by discovered source roots, tests, docs, scripts, and config paths instead.
Workflow Recommendation
MANDATORY IMPORTANT MUST ATTENTION — NO EXCEPTIONS: If you are NOT already in a workflow, use ask the user directly to ask the user:
- Execute
$release-doc directly (Recommended) — Standalone analysis and doc generation
- Run
$production-readiness-review + $quality-gate first — Pre-release quality gate, then release doc
Next Steps
MANDATORY IMPORTANT MUST ATTENTION — NO EXCEPTIONS after completing this skill, use ask the user directly to present:
- "$release-notes (Recommended)" — Generate consumer-facing release notes from the same range
- "$commit" — Commit the generated release doc to version control
- "$docs-update" — Sync project reference docs with any changes discovered during analysis
- "Skip, done" — User decides
[IMPORTANT] Use task tracking to break ALL work into small tasks BEFORE starting. This prevents context loss during large diff analysis. Always dump git artifacts to external files before reading.
External Memory: Write ALL intermediate findings and git artifacts to docs/release/ — this is MANDATORY. Large diffs will overflow context. Never try to hold a 30-day diff in memory.
Be skeptical. Apply critical thinking, sequential thinking. Every claim needs traced proof, confidence percentages (Idea should be more than 80%).
AI Mistake Prevention — Failure modes to avoid on every task:
Re-read files after context changes. Context compaction, resume, or long-running work can make memory stale; verify current files before acting.
Verify generated content against source evidence. AI hallucinates APIs, names, claims, and document facts. Check the relevant source before documenting or referencing.
Check downstream references before deleting or renaming. Removing an artifact can stale docs, generated mirrors, configs, and callers; map references first.
Trace the full impact chain after edits. Changing a definition can miss derived outputs and consumers. Follow the affected chain before declaring done.
Verify ALL affected outputs, not just the first. One green check is not all green checks; validate every output surface the change can affect.
Assume existing values are intentional — ask WHY before changing OR flagging one as a defect. Before changing or reporting a constant, limit, flag, cutoff, wording, or pattern, read nearby context and history, the CALLER's ordering, and 2+ sibling call sites of the same convention. A doc stating WHAT without WHY is missing rationale, not proof of a missing guard.
Surface ambiguity before acting — don't pick silently. Multiple valid interpretations require an explicit question or stated assumption with risk.
Assert the outcome your system owns, not the intermediate state your infrastructure owns. When verifying async work, assert the final business state — never the delivery/retry bookkeeping held in shared infrastructure that any co-running process can write. Such a check passes when run alone and flakes the moment anything else shares that infrastructure.
Keep shared guidance role-relevant. Universal guidance must help every receiving skill or agent; code-specific obligations belong only in code-specific protocols.
Critical Thinking Mindset — Apply critical thinking, sequential thinking. Every claim needs traced proof, confidence >80% to act.
Anti-hallucination: Never present guess as fact — cite sources for every claim, admit uncertainty freely, self-check output for errors, cross-reference independently, stay skeptical of own confidence — certainty without evidence root of all hallucination.
MUST ATTENTION apply critical + sequential thinking — every claim needs appropriate traced evidence (file:line for repo/code claims; source URL or artifact section for research, product, content, and docs claims); confidence >80% to act, <60% DO NOT recommend. Anti-hallucination: never present guess as fact, admit uncertainty freely, cross-reference independently, stay skeptical of own confidence.
MUST ATTENTION apply AI mistake prevention — verify generated content against evidence, trace downstream references before deleting or renaming, verify all affected outputs, re-read files after context loss, and surface ambiguity before acting.
Project Protocol Overlay — Before executing this skill, resolve any PROJECT overlay rules layered onto it: match this skill's name against the Target column of the project's skill-protocol index (docs/project-reference/skill-protocols-reference.md by default; a referenceDocs entry in docs/project-config.json overrides the path), taking the most specific matching tier ONLY — exact name > glob > *. That precedence orders overlays against EACH OTHER, never against this skill. Read ONLY the matched bodies, resolved as <protocols-dir>/<Name>.md; a row's Body link is display text, never a read path. A matched body that is missing or malformed is REPORTED and skipped — never reconstructed from the index Description. No index, or no match -> proceed with no overlay, silently. Full contract: .claude/skills/project-skill-protocol/references/registry.md.
Overlays are ADDITIVE ONLY: they ADD rules on top of this skill's own protocol and NEVER replace, override, disable, or reinterpret a rule it already states — removing every overlay must return this skill to exactly its documented behavior. An overlay is a BRIEF, not an authority escalation: it can NEVER waive a workflow gate, git discipline, a review gate, or a user-confirmation gate. A genuine overlay-vs-skill conflict, or two equally-specific overlays that directly contradict -> surface both to the user; NEVER resolve silently.
MUST ATTENTION resolve project protocol overlays for this skill BEFORE executing — most specific matching tier only (exact > glob > *, which ranks overlays against each other, NEVER against this skill), read only matched bodies at <protocols-dir>/<Name>.md; a missing or malformed body is reported, never reconstructed. Overlays are ADDITIVE ONLY (they never replace this skill's own rules) and are a brief, NEVER an authority escalation; an equal-specificity contradiction goes to the user.
Closing Reminders
Protocols in force (concise digest of the SYNC/shared blocks this skill carries):
- AI Mistake Prevention: verify generated content against evidence, trace downstream references, verify all affected outputs, re-read after context loss, surface ambiguity.
- Critical Thinking: traced
file:line proof per claim, confidence >80% to act, never guess.
IMPORTANT MUST ATTENTION dump ALL git artifacts to docs/release/ BEFORE reading any diff content
IMPORTANT MUST ATTENTION break work into small todo tasks using task tracking BEFORE starting
IMPORTANT MUST ATTENTION cite commit hash + file path for every claim in the generated doc
IMPORTANT MUST ATTENTION add a final review task to validate completeness against artifact files
[TASK-PLANNING] Before acting, analyze task scope and systematically break it into small todo tasks and sub-tasks using task tracking.
Hookless Prompt Protocol Mirror (Auto-Synced)
Source: .claude/.ck.json + .claude/skills/shared/sync-inline-versions.md (:full blocks) + .claude/scripts/lib/hookless-prompt-protocol.cjs
[WORKFLOW-EXECUTION-PROTOCOL] [BLOCKING] Workflow Execution Protocol — MANDATORY IMPORTANT MUST CRITICAL. Do not skip for any reason.
Generic portability boundary: Reusable skills and protocol text stay project-neutral; project-specific conventions are discovered from docs/project-config.json and docs/project-reference/. Apply shared AI-SDD from shared/sdd-artifact-contract.md. Read docs/project-config.json and docs/project-reference/docs-index-reference.md, then open the project reference docs named there. For spec, test-case, behavior-change, public-contract, or docs/specs/ work, route through the local spec docs named by the docs index: feature-spec-reference.md, spec-system-reference.md, spec-principles.md, and workflow-spec-test-code-cycle-reference.md when specs/tests/code must stay synchronized. If either file or a required reference doc is missing or stale, auto-run $project-init (or the narrow lower-level route such as $project-config, $docs-init, $scan-all, or $scan --target=<key>) before ordinary project-specific work. Any supported AI tool may execute when this shared context and local docs are available.
- DETECT: If the prompt starts with an explicit slash skill/workflow command, execute it directly. Otherwise match the prompt against the workflow catalog and skill list.
- ANALYZE: Choose the best option: execute directly, invoke a skill, activate a standard workflow, or compose a custom step combination.
- AUTO-SELECT: Pick the best option yourself. Do not ask the user to choose between direct execution, skill, standard workflow, or custom workflow.
- ACTIVATE: For a selected workflow, call
$start-workflow <workflowId>; for a selected skill, invoke that skill; for a custom workflow, sequence custom steps directly; for direct execution, proceed with the task.
- CREATE TASKS: task tracking for ALL workflow/skill/custom steps before execution when the selected path has multiple steps.
- PARALLELIZE: Before executing the task list, tag each task
PAR (independent inputs + write set disjoint from every other PAR task) or SEQ (name the blocking dependency), group PAR tasks into waves, declare the wave plan, and spawn each wave's sub-agents in ONE message — all-return barrier per wave, fan-out one level deep unless a sub-agent's own definition authorizes further fan-out. Sequential-by-default is a defect when tasks are independent; do not parallelize shared write targets, output-consuming tasks, trivial single-file work, ordering a skill or workflow explicitly fixes, or user-approval gates.
- EXECUTE: Advance per the Workflow Step Advancement & Parallel Phases rule in your context instructions — model-driven; a sub-agent completion advances a step identically to an inline call; a parallel-phase group is an all-return barrier (advance only after ALL members return, never serialize it)
Shared AI-SDD Protocol Markers
Source: .claude/skills/shared/sync-inline-versions.md
SYNC:ai-sdd-artifact-contract
AI-SDD Artifact Contract — Shared spec-driven development rules stay portable and source-owned.
- Keep reusable AI-SDD principles in
.claude; put repository-specific paths, commands, owners, products, and formats in project config/reference docs.
- Preserve cycle:
spec -> plan -> tasks -> implement -> verify -> update spec/docs.
- Trace every requirement or invariant through decision, task, TC/test, source evidence, and docs/spec update.
- Treat code-to-spec extraction as reference-only until accepted by the canonical spec owner.
- Any supported AI tool may plan, implement, review, or verify with synced context; using multiple tools is optional.
- Update
.claude source first, then sync generated mirrors; do not manually edit .agents, .codex, or AGENTS.md. — why: mirrors are generated artifacts; hand-edits are overwritten on the next sync
- If
docs/project-config.json, root instruction files, or a required project-reference doc is missing or stale, auto-run $project-init or the narrow lower-level route before ordinary project-specific work.
Active reference: shared/sdd-artifact-contract.md in the active skills root.
SYNC:ai-sdd-artifact-contract:reminder
- MANDATORY Apply
shared/sdd-artifact-contract.md; keep reusable AI-SDD in .claude and local rules in project docs.
- MANDATORY Code-to-spec extraction is reference-only until canonical acceptance; any supported AI tool may execute with synced context.
- MANDATORY Update
.claude source before syncing generated mirrors; do not manually edit .agents, .codex, or AGENTS.md.
- MANDATORY Missing or stale project config, root instruction files, or required reference docs route project-specific work through
$project-init or the narrow setup route automatically.
[TASK-PLANNING] [MANDATORY] BEFORE executing any workflow or skill step, create/update task tracking for all planned steps, then keep it synchronized as each step starts/completes.
[LESSON-LEARNED-REMINDER] [BLOCKING] Task Planning & Continuous Improvement — MANDATORY. Do not skip.
Break work into small tasks (task tracking) before starting. Add final task: "Analyze AI mistakes & lessons learned".
Extract lessons — ROOT CAUSE ONLY, not symptom fixes:
- Name the FAILURE MODE (reasoning/assumption failure), not symptom — "assumed API existed without reading source" not "used wrong enum value".
- Generality test: does this failure mode apply to ≥3 contexts/codebases? If not, abstract one level up.
- Write as a universal rule — strip project-specific names/paths/classes. Useful on any codebase.
- Consolidate: multiple mistakes sharing one failure mode → ONE lesson.
- Recurrence gate: "Would this recur in future session WITHOUT this reminder?" — No → skip
$learn.
- Auto-fix gate: "Could
$code-review/$code-simplifier/$security-review/$lint catch this?" — Yes → improve review skill instead.
- BOTH gates pass → ask user to run
$learn.
[CRITICAL-THINKING-MINDSET] Apply critical thinking, sequential thinking. Every claim needs traced proof, confidence >80% to act.
Anti-hallucination principle: Never present guess as fact — cite sources for every claim, admit uncertainty freely, self-check output for errors, cross-reference independently, stay skeptical of own confidence — certainty without evidence root of all hallucination.
AI Attention principle (Primacy-Recency): Put the 3 most critical rules at both top and bottom of long prompts/protocols so instruction adherence survives long context windows.
Goal-driven execution: Define success criteria first, loop until verified, and stop only when observable checks pass.
Tests verify intent: Tests must protect business rules/invariants and fail when the protected intent breaks, not only mirror current behavior.
Common AI Mistake Prevention (System Lessons)
- Re-read files after context compaction. Edit requires prior Read in same context; compaction wipes read state. Re-read before editing.
- Grep for old terms after bulk replacements. AI over-trusts find/replace completeness. Grep full repo after bulk edits for missed refs in docs/configs/catalogs.
- Check downstream references before deleting. Deletions cascade doc/code staleness. Map referencing files before removal.
- After memory loss, check existing state before creating new. Compaction wipes prior-work memory. Query current state to resume — never blindly duplicate.
- Verify AI-generated content against actual code. AI hallucinates APIs, class names, method signatures. Grep to confirm existence before documenting/referencing.
- Trace full dependency chain after edits. Changing a definition misses downstream consumers. Trace the full chain.
- When renaming, grep ALL consumer file types. Some file types silently ignore missing refs (no compile error). Search code, templates, configs, generated files.
- Trace ALL code paths when verifying correctness. Code existing ≠ code executing. Trace early exits, error branches, conditional skips — not just happy path.
- Update docs that embed canonical data when source changes. Docs inlining derived data (workflows, schemas, configs) go stale silently. Update all embedding docs alongside source.
- Verify sub-agent results after context recovery. Background agents may finish while parent compacted — grep-verify output, don't trust assumed completion.
- Cross-check full target list against sub-agent assignments. Parallel sub-agents by category miss boundary items. Reconcile union of assignments against target list before proceeding.
- Sub-agents inherit knowledge only from their agent .md definition — use custom agent types, not built-in Explore. Tool adoption = permission + knowledge + enforcement (numbered workflow step).
- Persist sub-agent findings incrementally, not as a final batch. Long sub-agents hit cutoffs before final write — findings lost. Instruct append-per-section to report file.
- When debugging, ask "whose responsibility?" before fixing. Trace caller (wrong data) vs callee (wrong handling). Fix at responsible layer — never patch symptom site.
- Test failure → record a provisional verdict before trace/edit, then investigate. Use the full five-way taxonomy: SOURCE-WRONG (production violates intent), TEST-WRONG (assertion/setup is stale), TEST-NOT-OPTIMAL (valid but fragile or low-signal test), ENVIRONMENT-BLOCKED (external state prevents a verdict), or AMBIGUOUS (intent/evidence cannot choose safely). Then trace root cause and triangulate against the governing spec (
docs/specs/** if one exists) AND source. NEVER weaken an assertion, add a skip, relax a timeout, or change source merely to force green.
- Grep ALL removed names after extraction/refactoring. Primary file "done" ≠ secondary files clean. Grep entire scope for every removed symbol before declaring complete.
- Assume existing values are intentional — ask WHY before changing OR flagging one as a defect. Pattern-matching as "wrong" skips context. Before changing or reporting any constant/limit/flag/cutoff: read comments, git blame, the CALLER's ordering (the guarantee that makes the value correct usually lives in code running immediately BEFORE the cited line), and 2+ sibling call sites of the same convention. A doc stating WHAT without WHY is missing rationale, not proof of a missing guard — and in a validation pass, an accurate
file:line citation proves the transcription, never the defect.
- Verify ALL affected outputs, not just the first. One build green ≠ all green. Multi-stack changes (backend/frontend/tests/docs) require verifying EVERY output.
- Evaluate fit before copying a nearby pattern. Closest example ≠ matching preconditions — verify the new context shares the same constraints, base classes, scope, lifetime.
- Holistic-first debugging — resist nearest-attention trap. Don't dive into first plausible cause. List EVERY precondition (config, env vars, paths, DB, endpoints, creds, versions, DI, data). Verify each against evidence (grep/query — not reasoning). Ask "what would falsify this?" — if nothing, it's not a hypothesis. Most expensive failure: going deeper in "obvious" layer while bug sits in layer never questioned.
- Surgical changes — apply the diff test (context-aware). Two modes: (1) Bug fix → every line traces to the bug; no restyling; orphan cleanup only for imports YOUR changes made unused. (2) Review/enhancement → implement improvements AND announce as "Enhancement beyond main request: [what]". Never silently scope-creep. Diff test: "Would this line exist if I wasn't asked to do X?" — if no, delete or announce.
- Surface ambiguity before coding — don't pick silently. Multiple valid interpretations → present each with effort: "[Request] could mean (1) [N h], (2) [N h]. Which matters?" List scope/format/volume/constraints assumptions first. If simpler path exists, say so. Never silently pick.
- [MANDATORY FIRST ACTION] ALWAYS activate a suitable skill or workflow BEFORE responding. Match task against workflow catalog + skill list; invoke via skill invocation or
$start-workflow <workflowId>. NEVER answer or write code before checking. Skip = protocol violation.
- Why-Review adversarial mindset — apply when reviewing any plan, decision, or design. Default SKEPTIC not VALIDATOR: steel-man a rejected alternative, invert each stated reason ("what does it sacrifice?"), stress-test top 2-3 assumptions, run pre-mortem ("ships, fails in 3 months — what breaks?"), surface 1-2 alternatives author missed. Section presence ≠ quality; quality = causal reasoning + concrete mitigations + evidence, not "it's better" or "monitor closely".
- Front-load report-write in sub-agent prompts for large reviews. Many-file sub-agents hit budget before final write — findings lost. Design prompts so: (1) report-write is first explicit deliverable, (2) append per-file/section (not batched), (3) scope bounded so reads don't exhaust budget. Truncated mid-sentence with no report file → spawn narrower scope, don't retry same prompt.
- After context compaction, re-verify all prior phase outcomes before continuing. Summaries describe intent, not environment state (git index, filesystem, processes). On resume, FIRST audit: git status, re-read modified files, verify filesystem. Every "completed" claim is an untested hypothesis until evidence confirms.
- OOM/memory: check row count before row size. Triage: (1) Unbounded query — no DB filter for trigger? Push filter to DB; eliminates OOM. (2) Large rows? Projection reduces proportionally. Row reduction > projection in ROI.
- Assert the outcome your system OWNS, never the intermediate state your INFRASTRUCTURE owns. When testing anything asynchronous (queue/broker delivery, retries, background jobs, caches, replication), assert the final business/entity state. NEVER assert the delivery bookkeeping — consume/send status, attempt counts, last-error, row existence or counts in a broker, scheduler, or outbox/inbox table. That bookkeeping lives in shared infrastructure that ANY co-running process (a peer worker, a second replica, a leftover local container) can write, usually under a deterministic shared key, so the assertion silently tests the developer's environment instead of the system: green when run alone, flaky the instant anything else shares that broker + database. Gate question for every assertion: "would this hold no matter WHICH process did the work?" — if no, assert the converged data state instead. Corollary: process-local fault injection and in-process telemetry cannot gate work any process may perform — use them as stress amplifiers (arm → bounded window → disarm → assert convergence), never as preconditions.
- Keep domain concepts out of generic/shared/infrastructure layers. Reusable layer (shared library, framework, infra module) must reference NO consumer-specific domain concept — tenant/customer/product IDs, business entities, feature rules. Leak compiles + runs → passes review silently while coupling the "reusable" layer to one consumer. Keep shared type domain-free; push domain fields/logic down into the consumer via subclass/composition. — why: a layer coupled to one consumer's domain is no longer reusable.
1---
2name: release-doc
3description: '[Documentation] Use when you need to generate a detailed, AI-analyzed release document from git history over a time range or custom prompt.'
4metadata:
5 source: https://skillmd.com/skills/duc01226/release-doc
6---
7
8> Codex compatibility note:
9>
10> - Invoke repository skills with `$skill-name` in Codex; this mirrored copy rewrites legacy Claude `/skill-name` references.
11> - Task tracker mandate: BEFORE executing any workflow or skill step, create/update task tracking for all steps and keep it synchronized as progress changes.
12> - User-question prompts mean to ask the user directly in Codex.
13> - Ignore Claude-specific mode-switch instructions when they appear.
14> - Strict execution contract: when a user explicitly invokes a skill, execute that skill protocol as written.
15> - Subagent authorization: when a skill is user-invoked or AI-detected and its protocol requires subagents, that skill activation authorizes use of the required `spawn_agent` subagent(s) for that task.
16> - Do not skip, reorder, or merge protocol steps unless the user explicitly approves the deviation first.
17> - For workflow skills, execute each listed child-skill step explicitly and report step-by-step evidence.
18> - If a required step/tool cannot run in this environment, stop and ask the user before adapting.
19
20<!-- CODEX:PROJECT-REFERENCE-LOADING:START -->
21
22## Codex Project-Reference Loading (No Hooks)
23
24Codex uses static project-reference loading instead of runtime-injected project docs.
25When coding, planning, debugging, testing, or reviewing, open project docs explicitly using this routing.
26
27**Always read:**
28
29- `docs/project-config.json` (project-specific paths, commands, modules, and workflow/test settings)
30- `docs/project-reference/docs-index-reference.md` (routes to the full `docs/project-reference/*` catalog)
31- `docs/project-reference/lessons.md` (always-on guardrails and anti-patterns)
32
33**Missing/stale context route:** If `docs/project-config.json`, the docs index, `lessons.md`, `CLAUDE.md`, `AGENTS.md`, or any task-required reference doc is missing or stale, auto-run `$project-init` or the narrow setup route (`$project-config`, `$docs-init`, `$scan-all`, `$scan --target=<key>`, `$claude-md-init`) before ordinary project-specific work. If Codex mirrors or `AGENTS.md` are missing/stale, ask the user to run `$sync-codex`; do not auto-run it.
34
35**Situation-based docs:**
36
37- Project structure/architecture/tech-stack/deployment/setup (any layer — backend, frontend, or infra): `project-structure-reference.md`
38- Backend/CQRS/API/domain/entity changes: `backend-patterns-reference.md`, `domain-entities-reference.md`
39- Frontend/UI/styling/design-system: `frontend-patterns-reference.md`, `scss-styling-guide.md`, `design-system/README.md`
40- Spec authoring, `docs/specs/` pathing, or TC format: `feature-spec-reference.md`, `spec-system-reference.md`, `spec-principles.md`
41- Behavior/public-contract changes or spec-test-code sync: `workflow-spec-test-code-cycle-reference.md` plus the spec docs above
42- Derived spec indexes/ERDs/reimplementation guides: `spec-system-reference.md` and source Feature Specs under `docs/specs/`
43- Integration test implementation/review: `integration-test-reference.md`
44- E2E test implementation/review: `e2e-test-reference.md`
45- Code review/audit work: `code-review-rules.md` plus domain docs above based on changed files
46
47Do not read all docs blindly. Start from `docs-index-reference.md`, then open only relevant files for the task.
48
49<!-- CODEX:PROJECT-REFERENCE-LOADING:END -->
50
51## Quick Summary
52
53**Goal:** Generate a detailed, AI-analyzed release document from git history — by time range (e.g., last 30 days) or custom comparison (e.g., `v1.0.0..HEAD`) or a user-described focus area.
54
55**Workflow:**
56
571. **Resolve Range** — Determine the git range from user input
582. **Dump Artifacts** — Save log, file-status, and full diff to `docs/release/` BEFORE analyzing
593. **Categorize Changes** — Group changed files by framework area (hooks, skills, agents, workflows, principles, config, docs)
604. **Read Key Diffs** — For each category, read the most significant changes via `git show` or `git diff`
615. **Generate Doc** — Write structured markdown release doc to `docs/release/`
626. **Validate** — Verify completeness and accuracy against the artifact files
63
64**Key Rules:**
65
66- **Dump first, read second** — NEVER analyze a diff you haven't saved to a file first
67- **Group by theme** — Do not list commits chronologically; group by what area they impact
68- **Cite evidence** — Every section claim must reference a specific commit hash or file path
69- **Custom focus honored** — If user provides a focus prompt, weight that area more deeply
70- **Draft status** — All generated docs are Draft; add "Status: Draft" to the header
71
72---
73
74# Release Doc Skill
75
76Generate a detailed, narrative release document by analyzing git history over a time range or between two refs.
77
78## Invocation
79
80```
81$release-doc [--days N] [--since DATE] [--range base..head] [--focus "custom prompt"] [--output path]
82```
83
84**Examples:**
85
86```bash
87# Last 30 days
88$release-doc --days 30
89
90# Since a specific date
91$release-doc --since 2026-03-15
92
93# Between two tags or commits
94$release-doc --range v1.0.0..HEAD
95
96# With custom focus area
97$release-doc --days 30 --focus "what changed in hooks and workflow enforcement"
98
99# Specify output file
100$release-doc --days 30 --output docs/release/release-2026-Q1.md
101```
102
103## Differences from `$release-notes` and `$changelog`
104
105| Feature | `$release-doc` | `$release-notes` | `$changelog` |
106| ------------------------ | ---------------------------- | ------------------------- | ------------------- |
107| Time-based ranges | ✅ `--days N` / `--since` | ❌ ref-to-ref only | ❌ PR/commit only |
108| AI thematic analysis | ✅ Groups by framework area | ❌ Commit-type categories | ❌ File-by-file |
109| Dumps artifacts to files | ✅ Mandatory before analysis | ❌ Pipes in memory | ❌ In memory |
110| Custom focus prompt | ✅ `--focus "..."` | ❌ | ❌ |
111| Narrative sections | ✅ Context + rationale | ❌ Bullet list only | ❌ Bullet list only |
112| Framework-level docs | ✅ hooks/skills/agents | ❌ App features | ❌ App features |
113
114## Step 1: Resolve Range
115
116Determine the git range from the user's input:
117
118```bash
119# Time-based (--days N or --since DATE)
120SINCE_DATE=$(date -d "-30 days" +%Y-%m-%d) # Linux
121SINCE_DATE=$(date -v-30d +%Y-%m-%d) # macOS
122git log --since="$SINCE_DATE" --oneline --format="%H %ad %s" --date=short
123
124# Ref-based (--range base..head)
125git log v1.0.0..HEAD --oneline --format="%H %ad %s" --date=short
126
127# Count commits in range
128git log --since="$SINCE_DATE" --oneline | wc -l
129```
130
131**Identify the boundary commits:**
132
133```bash
134# Oldest commit in range (for diff base)
135OLDEST=$(git log --since="$SINCE_DATE" --format="%H" | tail -1)
136# Newest
137NEWEST=HEAD
138```
139
140## Step 2: Dump Git Artifacts (MANDATORY BEFORE ANALYSIS)
141
142**CRITICAL:** Run ALL dumps before reading ANY content. Use background runs for large diffs.
143
144```bash
145# Create output directory
146mkdir -p docs/release
147
148# 1. Full git log with bodies
149git log --since="$SINCE_DATE" --format="%H %ad %s%n%b" --date=short \
150 > docs/release/git-log-{PERIOD}.txt
151
152# 2. File-level change status (A/M/D)
153git diff ${OLDEST}^..HEAD --name-status \
154 > docs/release/diff-file-status-{PERIOD}.txt
155
156# 3. Diff stat summary (line counts)
157git diff ${OLDEST}^..HEAD --stat \
158 > docs/release/diff-stat-{PERIOD}.txt
159
160# 4. Full consolidated diff (background — may be large)
161git diff ${OLDEST}^..HEAD \
162 > docs/release/git-diff-{PERIOD}-full.txt &
163
164echo "Artifacts saved. Commit count: $(cat docs/release/git-log-{PERIOD}.txt | grep -c '^[a-f0-9]\{40\}')"
165```
166
167Replace `{PERIOD}` with a human-readable period string, e.g., `30d`, `2026-03-15-to-2026-04-14`.
168
169**Verify all artifact files exist and are non-empty before proceeding.**
170
171## Step 3: Categorize Changes by Framework Area
172
173Read `docs/release/diff-file-status-{PERIOD}.txt` and group files into categories:
174
175```bash
176# Count by area
177grep -c "^[AMD]\s*.claude/hooks/" docs/release/diff-file-status-{PERIOD}.txt
178grep -c "^[AMD]\s*.claude/skills/" docs/release/diff-file-status-{PERIOD}.txt
179grep -c "^[AMD]\s*.claude/agents/" docs/release/diff-file-status-{PERIOD}.txt
180grep -c "^[AMD]\s*.claude/workflows" docs/release/diff-file-status-{PERIOD}.txt
181grep -c "^[AMD]\s*.claude/docs/" docs/release/diff-file-status-{PERIOD}.txt
182grep -c "^[AMD]\s*docs/" docs/release/diff-file-status-{PERIOD}.txt
183grep -c "^[AMD]\s*CLAUDE.md" docs/release/diff-file-status-{PERIOD}.txt
184```
185
186**Standard category map for easy-claude projects:**
187
188| File Path Pattern | Category |
189| ------------------------------------ | ----------------------- |
190| `.claude/hooks/**` | Hook Enhancements |
191| `.claude/hooks/lib/**` | Hook Library |
192| `.claude/skills/**` | Skills |
193| `.claude/agents/**` | Agent Definitions |
194| `.claude/workflows/**` | Workflow Orchestration |
195| `.claude/workflows.json` | Workflow Registry |
196| `.claude/docs/**` | Framework Documentation |
197| `.claude/scripts/**` | Tooling & Scripts |
198| `docs/project-reference/**` | Project Reference Docs |
199| `CLAUDE.md` | Principles & Core Rules |
200| `.claude/.ck.json` / `settings.json` | Configuration |
201
202**For non-easy-claude projects**, derive categories from the file path patterns actually present.
203
204## Step 4: Analyze Key Changes Per Category
205
206For each category with significant changes (>5 files or >200 lines), read representative diffs:
207
208```bash
209# Most impactful commits (by files changed)
210git log --since="$SINCE_DATE" --format="%H %s" --name-only | \
211 grep -A100 "^[a-f0-9]\{40\}" | head -200
212
213# Show specific commit detail
214git show {COMMIT_HASH} --stat --format="%s%n%b"
215
216# Diff for a specific file across the range
217git diff ${OLDEST}^..HEAD -- .claude/hooks/init-prompt-gate.cjs
218```
219
220**Analysis questions for each category:**
221
222- **Hooks:** What new gates or enforcement were added? What context injection changed? What was the trigger behavior before vs after?
223- **Skills:** What new skills were created? What protocols were updated in existing skills? What was removed or merged?
224- **Agents:** What agent definitions changed? Did tool lists or dispatch rules change?
225- **Workflows:** What workflow steps were added/removed? Did sequence order change? Were any workflows merged or deleted?
226- **Principles:** What rules in CLAUDE.md changed? Were confidence thresholds tightened? New hard gates added?
227- **Config:** What new schema fields? What validation was added?
228- **Scripts/Tooling:** What new utilities? What improved or automated?
229
230**Evidence requirement:** For every claim in the doc, record `{commit_hash}:{file_path}` as your source.
231
232## Step 5: Handle Custom Focus Prompt
233
234If user provided `--focus "..."`, apply additional depth to that area:
235
2361. Grep all changed files matching the focus keywords
2372. Read the full diff for matching files (not just stat)
2383. Include a dedicated top-level section for the focus area in the output doc
2394. Cross-reference related changes in other areas (e.g., a new skill + its hook + its workflow entry)
240
241## Step 6: Generate the Release Doc
242
243Write to `docs/release/release-notes-{PERIOD}.md` (or `--output` path).
244
245**Required structure:**
246
247```markdown
248# Release Notes — {Project Name}
249
250**Period:** {START} → {END}
251**Commits:** {N} commits | **Files changed:** {N} | **Insertions:** +{N} | **Deletions:** −{N}
252**Status:** Draft
253
254---
255
256## Executive Summary
257
2583-4 sentences. The most transformative changes, their business/developer impact, and the overall trajectory.
259
260---
261
262## {Category 1}: {Theme Name}
263
264### What Changed
265
266{Narrative description — not bullet-list of commits}
267
268### Why It Matters
269
270{Impact on how developers/AI uses the system}
271
272### Key Additions / Modifications
273
274| Item | Type | Effect |
275| ---- | ---- | ------ |
276
277---
278
279## {Category 2}: ...
280
281[repeat for each significant category]
282
283---
284
285## New Skills Added
286
287| Skill | Command | Purpose |
288| ----- | ------- | ------- |
289
290## Skills Removed or Merged
291
292| Skill | Reason |
293| ----- | ------ |
294
295## Principles & Rules Updated
296
297| Rule | Before | After | Reason |
298| ---- | ------ | ----- | ------ |
299
300---
301
302## Summary Statistics
303
304| Category | Count |
305| -------------- | ----- |
306| Commits | {N} |
307| Files changed | {N} |
308| New skills | {N} |
309| Skills removed | {N} |
310| New hooks | {N} |
311| Hooks modified | {N} |
312| Agents updated | {N} |
313| Net lines | +{N} |
314
315---
316
317_Generated: {DATE} | Branch: {BRANCH} | Range: {OLDEST_HASH}..{HEAD_HASH}_
318_Artifacts: `docs/release/git-log-{PERIOD}.txt`, `docs/release/diff-file-status-{PERIOD}.txt`_
319```
320
321## Step 7: Validate
322
323After writing, verify:
324
325- [ ] Every category section references at least one commit hash
326- [ ] "New Skills" table is complete (cross-check against `grep "^A.*skills" diff-file-status-*.txt`)
327- [ ] "Skills Removed" is complete (cross-check against `grep "^D.*skills" diff-file-status-*.txt`)
328- [ ] Summary Statistics match the actual diff stat
329- [ ] No section says "TODO" or contains unfilled template placeholders
330- [ ] Artifact files are listed in the doc footer
331
332## Configuration: Category Map Override
333
334For projects with custom directory structures, override the category map via `project-config.json`:
335
336```json
337{
338 "releaseDoc": {
339 "categoryMap": {
340 "{api-source-root}/**": "API Layer",
341 "{ui-source-root}/**": "Frontend",
342 "migrations/**": "Database Schema"
343 },
344 "outputDir": "docs/release",
345 "focusAreas": ["hooks", "skills", "agents"]
346 }
347}
348```
349
350## Integration with Other Skills
351
352- **`$release-notes`** — Use after `$release-doc` to generate a consumer-facing version (tag-to-tag, conventional commits)
353- **`$changelog`** — Use for individual feature changelog entries; `$release-doc` is for multi-week summaries
354- **`$graph-blast-radius`** — Run before release to assess impact of the most changed files
355- **`$docs-update`** — Update project reference docs after analyzing what changed
356- **`$commit`** — Commit the generated release doc
357
358## Troubleshooting
359
360### Diff too large for context
361
362**Solution:** The skill requires artifact files be written first — never read the full diff into context. Use `git show {hash}` for individual commits and `grep -l` to narrow file scope.
363
364### Time range yields 0 commits
365
366```bash
367# Verify date format (ISO 8601)
368git log --since="2026-03-15" --oneline | head -5
369# If empty, check git's date interpretation:
370git log --format="%ad" --date=short | head -5
371```
372
373### Non-easy-claude project structure
374
375The skill auto-derives categories from file paths. For repositories without `.claude/`, group by discovered source roots, tests, docs, scripts, and config paths instead.
376
377---
378
379## Workflow Recommendation
380
381> **MANDATORY IMPORTANT MUST ATTENTION — NO EXCEPTIONS:** If you are NOT already in a workflow, use ask the user directly to ask the user:
382>
383> 1. **Execute `$release-doc` directly (Recommended)** — Standalone analysis and doc generation
384> 2. **Run `$production-readiness-review` + `$quality-gate` first** — Pre-release quality gate, then release doc
385
386---
387
388## Next Steps
389
390**MANDATORY IMPORTANT MUST ATTENTION — NO EXCEPTIONS** after completing this skill, use ask the user directly to present:
391
392- **"$release-notes (Recommended)"** — Generate consumer-facing release notes from the same range
393- **"$commit"** — Commit the generated release doc to version control
394- **"$docs-update"** — Sync project reference docs with any changes discovered during analysis
395- **"Skip, done"** — User decides
396
397---
398
399> **[IMPORTANT]** Use task tracking to break ALL work into small tasks BEFORE starting. This prevents context loss during large diff analysis. Always dump git artifacts to external files before reading.
400
401> **External Memory:** Write ALL intermediate findings and git artifacts to `docs/release/` — this is MANDATORY. Large diffs will overflow context. Never try to hold a 30-day diff in memory.
402
403**Be skeptical. Apply critical thinking, sequential thinking. Every claim needs traced proof, confidence percentages (Idea should be more than 80%).**
404
405<!-- SYNC:ai-mistake-prevention -->
406
407> **AI Mistake Prevention** — Failure modes to avoid on every task:
408>
409> **Re-read files after context changes.** Context compaction, resume, or long-running work can make memory stale; verify current files before acting.
410> **Verify generated content against source evidence.** AI hallucinates APIs, names, claims, and document facts. Check the relevant source before documenting or referencing.
411> **Check downstream references before deleting or renaming.** Removing an artifact can stale docs, generated mirrors, configs, and callers; map references first.
412> **Trace the full impact chain after edits.** Changing a definition can miss derived outputs and consumers. Follow the affected chain before declaring done.
413> **Verify ALL affected outputs, not just the first.** One green check is not all green checks; validate every output surface the change can affect.
414> **Assume existing values are intentional — ask WHY before changing OR flagging one as a defect.** Before changing or reporting a constant, limit, flag, cutoff, wording, or pattern, read nearby context and history, the CALLER's ordering, and 2+ sibling call sites of the same convention. A doc stating WHAT without WHY is missing rationale, not proof of a missing guard.
415> **Surface ambiguity before acting — don't pick silently.** Multiple valid interpretations require an explicit question or stated assumption with risk.
416> **Assert the outcome your system owns, not the intermediate state your infrastructure owns.** When verifying async work, assert the final business state — never the delivery/retry bookkeeping held in shared infrastructure that any co-running process can write. Such a check passes when run alone and flakes the moment anything else shares that infrastructure.
417> **Keep shared guidance role-relevant.** Universal guidance must help every receiving skill or agent; code-specific obligations belong only in code-specific protocols.
418
419<!-- /SYNC:ai-mistake-prevention -->
420
421<!-- SYNC:critical-thinking-mindset -->
422
423> **Critical Thinking Mindset** — Apply critical thinking, sequential thinking. Every claim needs traced proof, confidence >80% to act.
424> **Anti-hallucination:** Never present guess as fact — cite sources for every claim, admit uncertainty freely, self-check output for errors, cross-reference independently, stay skeptical of own confidence — certainty without evidence root of all hallucination.
425
426<!-- /SYNC:critical-thinking-mindset -->
427
428<!-- SYNC:critical-thinking-mindset:reminder -->
429
430**MUST ATTENTION** apply critical + sequential thinking — every claim needs appropriate traced evidence (`file:line` for repo/code claims; source URL or artifact section for research, product, content, and docs claims); confidence >80% to act, <60% DO NOT recommend. Anti-hallucination: never present guess as fact, admit uncertainty freely, cross-reference independently, stay skeptical of own confidence.
431
432<!-- /SYNC:critical-thinking-mindset:reminder -->
433
434<!-- SYNC:ai-mistake-prevention:reminder -->
435
436**MUST ATTENTION** apply AI mistake prevention — verify generated content against evidence, trace downstream references before deleting or renaming, verify all affected outputs, re-read files after context loss, and surface ambiguity before acting.
437
438<!-- /SYNC:ai-mistake-prevention:reminder -->
439
440<!-- SYNC:project-protocol-overlay -->
441
442> **Project Protocol Overlay** — Before executing this skill, resolve any PROJECT overlay rules layered onto it: match this skill's name against the `Target` column of the project's skill-protocol index (`docs/project-reference/skill-protocols-reference.md` by default; a `referenceDocs` entry in `docs/project-config.json` overrides the path), taking the most specific matching tier ONLY — exact name > glob > `*`. **That precedence orders overlays against EACH OTHER, never against this skill.** Read ONLY the matched bodies, resolved as `<protocols-dir>/<Name>.md`; a row's Body link is display text, never a read path. A matched body that is missing or malformed is REPORTED and skipped — never reconstructed from the index Description. No index, or no match -> proceed with no overlay, silently. Full contract: `.claude/skills/project-skill-protocol/references/registry.md`.
443>
444> Overlays are **ADDITIVE ONLY**: they ADD rules on top of this skill's own protocol and NEVER replace, override, disable, or reinterpret a rule it already states — removing every overlay must return this skill to exactly its documented behavior. An overlay is a BRIEF, not an authority escalation: it can NEVER waive a workflow gate, git discipline, a review gate, or a user-confirmation gate. A genuine overlay-vs-skill conflict, or two equally-specific overlays that directly contradict -> surface both to the user; NEVER resolve silently.
445
446<!-- /SYNC:project-protocol-overlay -->
447
448<!-- SYNC:project-protocol-overlay:reminder -->
449
450**MUST ATTENTION** resolve project protocol overlays for this skill BEFORE executing — most specific matching tier only (exact > glob > `*`, which ranks overlays against each other, NEVER against this skill), read only matched bodies at `<protocols-dir>/<Name>.md`; a missing or malformed body is reported, never reconstructed. Overlays are ADDITIVE ONLY (they never replace this skill's own rules) and are a brief, NEVER an authority escalation; an equal-specificity contradiction goes to the user.
451
452<!-- /SYNC:project-protocol-overlay:reminder -->
453
454## Closing Reminders
455
456**Protocols in force (concise digest of the SYNC/shared blocks this skill carries):**
457
458- **AI Mistake Prevention:** verify generated content against evidence, trace downstream references, verify all affected outputs, re-read after context loss, surface ambiguity.
459- **Critical Thinking:** traced `file:line` proof per claim, confidence >80% to act, never guess.
460
461**IMPORTANT MUST ATTENTION** dump ALL git artifacts to `docs/release/` BEFORE reading any diff content
462**IMPORTANT MUST ATTENTION** break work into small todo tasks using task tracking BEFORE starting
463**IMPORTANT MUST ATTENTION** cite commit hash + file path for every claim in the generated doc
464**IMPORTANT MUST ATTENTION** add a final review task to validate completeness against artifact files
465
466**[TASK-PLANNING]** Before acting, analyze task scope and systematically break it into small todo tasks and sub-tasks using task tracking.
467
468<!-- CODEX:SYNC-PROMPT-PROTOCOLS:START -->
469
470## Hookless Prompt Protocol Mirror (Auto-Synced)
471
472Source: `.claude/.ck.json` + `.claude/skills/shared/sync-inline-versions.md` (`:full` blocks) + `.claude/scripts/lib/hookless-prompt-protocol.cjs`
473
474## [WORKFLOW-EXECUTION-PROTOCOL] [BLOCKING] Workflow Execution Protocol — MANDATORY IMPORTANT MUST CRITICAL. Do not skip for any reason.
475
476**Generic portability boundary:** Reusable skills and protocol text stay project-neutral; project-specific conventions are discovered from docs/project-config.json and docs/project-reference/. Apply shared AI-SDD from `shared/sdd-artifact-contract.md`. Read `docs/project-config.json` and `docs/project-reference/docs-index-reference.md`, then open the project reference docs named there. For spec, test-case, behavior-change, public-contract, or `docs/specs/` work, route through the local spec docs named by the docs index: `feature-spec-reference.md`, `spec-system-reference.md`, `spec-principles.md`, and `workflow-spec-test-code-cycle-reference.md` when specs/tests/code must stay synchronized. If either file or a required reference doc is missing or stale, auto-run `$project-init` (or the narrow lower-level route such as `$project-config`, `$docs-init`, `$scan-all`, or `$scan --target=<key>`) before ordinary project-specific work. Any supported AI tool may execute when this shared context and local docs are available.
477
4781. **DETECT:** If the prompt starts with an explicit slash skill/workflow command, execute it directly. Otherwise match the prompt against the workflow catalog and skill list.
4792. **ANALYZE:** Choose the best option: execute directly, invoke a skill, activate a standard workflow, or compose a custom step combination.
4803. **AUTO-SELECT:** Pick the best option yourself. Do not ask the user to choose between direct execution, skill, standard workflow, or custom workflow.
4814. **ACTIVATE:** For a selected workflow, call `$start-workflow <workflowId>`; for a selected skill, invoke that skill; for a custom workflow, sequence custom steps directly; for direct execution, proceed with the task.
4825. **CREATE TASKS:** task tracking for ALL workflow/skill/custom steps before execution when the selected path has multiple steps.
4836. **PARALLELIZE:** Before executing the task list, tag each task `PAR` (independent inputs + write set disjoint from every other `PAR` task) or `SEQ` (name the blocking dependency), group `PAR` tasks into waves, declare the wave plan, and spawn each wave's sub-agents in ONE message — all-return barrier per wave, fan-out one level deep unless a sub-agent's own definition authorizes further fan-out. Sequential-by-default is a defect when tasks are independent; do not parallelize shared write targets, output-consuming tasks, trivial single-file work, ordering a skill or workflow explicitly fixes, or user-approval gates.
4847. **EXECUTE:** Advance per the **Workflow Step Advancement & Parallel Phases** rule in your context instructions — model-driven; a sub-agent completion advances a step identically to an inline call; a parallel-phase group is an all-return barrier (advance only after ALL members return, never serialize it)
485
486## Shared AI-SDD Protocol Markers
487
488Source: `.claude/skills/shared/sync-inline-versions.md`
489
490## SYNC:ai-sdd-artifact-contract
491
492> **AI-SDD Artifact Contract** — Shared spec-driven development rules stay portable and source-owned.
493>
494> 1. Keep reusable AI-SDD principles in `.claude`; put repository-specific paths, commands, owners, products, and formats in project config/reference docs.
495> 2. Preserve cycle: `spec -> plan -> tasks -> implement -> verify -> update spec/docs`.
496> 3. Trace every requirement or invariant through decision, task, TC/test, source evidence, and docs/spec update.
497> 4. Treat code-to-spec extraction as reference-only until accepted by the canonical spec owner.
498> 5. Any supported AI tool may plan, implement, review, or verify with synced context; using multiple tools is optional.
499> 6. Update `.claude` source first, then sync generated mirrors; do not manually edit `.agents`, `.codex`, or `AGENTS.md`. — why: mirrors are generated artifacts; hand-edits are overwritten on the next sync
500> 7. If `docs/project-config.json`, root instruction files, or a required project-reference doc is missing or stale, auto-run `$project-init` or the narrow lower-level route before ordinary project-specific work.
501>
502> **Active reference:** `shared/sdd-artifact-contract.md` in the active skills root.
503
504---
505
506## SYNC:ai-sdd-artifact-contract:reminder
507
508- **MANDATORY** Apply `shared/sdd-artifact-contract.md`; keep reusable AI-SDD in `.claude` and local rules in project docs.
509- **MANDATORY** Code-to-spec extraction is reference-only until canonical acceptance; any supported AI tool may execute with synced context.
510- **MANDATORY** Update `.claude` source before syncing generated mirrors; do not manually edit `.agents`, `.codex`, or `AGENTS.md`.
511- **MANDATORY** Missing or stale project config, root instruction files, or required reference docs route project-specific work through `$project-init` or the narrow setup route automatically.
512 **[TASK-PLANNING] [MANDATORY]** BEFORE executing any workflow or skill step, create/update task tracking for all planned steps, then keep it synchronized as each step starts/completes.
513
514## [LESSON-LEARNED-REMINDER] [BLOCKING] Task Planning & Continuous Improvement — MANDATORY. Do not skip.
515
516Break work into small tasks (task tracking) before starting. Add final task: "Analyze AI mistakes & lessons learned".
517
518**Extract lessons — ROOT CAUSE ONLY, not symptom fixes:**
519
5201. Name the FAILURE MODE (reasoning/assumption failure), not symptom — "assumed API existed without reading source" not "used wrong enum value".
5212. Generality test: does this failure mode apply to ≥3 contexts/codebases? If not, abstract one level up.
5223. Write as a universal rule — strip project-specific names/paths/classes. Useful on any codebase.
5234. Consolidate: multiple mistakes sharing one failure mode → ONE lesson.
5245. **Recurrence gate:** "Would this recur in future session WITHOUT this reminder?" — No → skip `$learn`.
5256. **Auto-fix gate:** "Could `$code-review`/`$code-simplifier`/`$security-review`/`$lint` catch this?" — Yes → improve review skill instead.
5267. BOTH gates pass → ask user to run `$learn`.
527 **[CRITICAL-THINKING-MINDSET]** Apply critical thinking, sequential thinking. Every claim needs traced proof, confidence >80% to act.
528 **Anti-hallucination principle:** Never present guess as fact — cite sources for every claim, admit uncertainty freely, self-check output for errors, cross-reference independently, stay skeptical of own confidence — certainty without evidence root of all hallucination.
529 **AI Attention principle (Primacy-Recency):** Put the 3 most critical rules at both top and bottom of long prompts/protocols so instruction adherence survives long context windows.
530 **Goal-driven execution:** Define success criteria first, loop until verified, and stop only when observable checks pass.
531 **Tests verify intent:** Tests must protect business rules/invariants and fail when the protected intent breaks, not only mirror current behavior.
532
533## Common AI Mistake Prevention (System Lessons)
534
535- **Re-read files after context compaction.** Edit requires prior Read in same context; compaction wipes read state. Re-read before editing.
536- **Grep for old terms after bulk replacements.** AI over-trusts find/replace completeness. Grep full repo after bulk edits for missed refs in docs/configs/catalogs.
537- **Check downstream references before deleting.** Deletions cascade doc/code staleness. Map referencing files before removal.
538- **After memory loss, check existing state before creating new.** Compaction wipes prior-work memory. Query current state to resume — never blindly duplicate.
539- **Verify AI-generated content against actual code.** AI hallucinates APIs, class names, method signatures. Grep to confirm existence before documenting/referencing.
540- **Trace full dependency chain after edits.** Changing a definition misses downstream consumers. Trace the full chain.
541- **When renaming, grep ALL consumer file types.** Some file types silently ignore missing refs (no compile error). Search code, templates, configs, generated files.
542- **Trace ALL code paths when verifying correctness.** Code existing ≠ code executing. Trace early exits, error branches, conditional skips — not just happy path.
543- **Update docs that embed canonical data when source changes.** Docs inlining derived data (workflows, schemas, configs) go stale silently. Update all embedding docs alongside source.
544- **Verify sub-agent results after context recovery.** Background agents may finish while parent compacted — grep-verify output, don't trust assumed completion.
545- **Cross-check full target list against sub-agent assignments.** Parallel sub-agents by category miss boundary items. Reconcile union of assignments against target list before proceeding.
546- **Sub-agents inherit knowledge only from their agent .md definition — use custom agent types, not built-in Explore.** Tool adoption = permission + knowledge + enforcement (numbered workflow step).
547- **Persist sub-agent findings incrementally, not as a final batch.** Long sub-agents hit cutoffs before final write — findings lost. Instruct append-per-section to report file.
548- **When debugging, ask "whose responsibility?" before fixing.** Trace caller (wrong data) vs callee (wrong handling). Fix at responsible layer — never patch symptom site.
549- **Test failure → record a provisional verdict before trace/edit, then investigate.** Use the full five-way taxonomy: SOURCE-WRONG (production violates intent), TEST-WRONG (assertion/setup is stale), TEST-NOT-OPTIMAL (valid but fragile or low-signal test), ENVIRONMENT-BLOCKED (external state prevents a verdict), or AMBIGUOUS (intent/evidence cannot choose safely). Then trace root cause and triangulate against the governing spec (`docs/specs/**` if one exists) AND source. NEVER weaken an assertion, add a skip, relax a timeout, or change source merely to force green.
550- **Grep ALL removed names after extraction/refactoring.** Primary file "done" ≠ secondary files clean. Grep entire scope for every removed symbol before declaring complete.
551- **Assume existing values are intentional — ask WHY before changing OR flagging one as a defect.** Pattern-matching as "wrong" skips context. Before changing or reporting any constant/limit/flag/cutoff: read comments, git blame, the CALLER's ordering (the guarantee that makes the value correct usually lives in code running immediately BEFORE the cited line), and 2+ sibling call sites of the same convention. A doc stating WHAT without WHY is missing rationale, not proof of a missing guard — and in a validation pass, an accurate `file:line` citation proves the transcription, never the defect.
552- **Verify ALL affected outputs, not just the first.** One build green ≠ all green. Multi-stack changes (backend/frontend/tests/docs) require verifying EVERY output.
553- **Evaluate fit before copying a nearby pattern.** Closest example ≠ matching preconditions — verify the new context shares the same constraints, base classes, scope, lifetime.
554- **Holistic-first debugging — resist nearest-attention trap.** Don't dive into first plausible cause. List EVERY precondition (config, env vars, paths, DB, endpoints, creds, versions, DI, data). Verify each against evidence (grep/query — not reasoning). Ask "what would falsify this?" — if nothing, it's not a hypothesis. Most expensive failure: going deeper in "obvious" layer while bug sits in layer never questioned.
555- **Surgical changes — apply the diff test (context-aware).** Two modes: (1) Bug fix → every line traces to the bug; no restyling; orphan cleanup only for imports YOUR changes made unused. (2) Review/enhancement → implement improvements AND announce as "Enhancement beyond main request: [what]". Never silently scope-creep. Diff test: "Would this line exist if I wasn't asked to do X?" — if no, delete or announce.
556- **Surface ambiguity before coding — don't pick silently.** Multiple valid interpretations → present each with effort: "[Request] could mean (1) [N h], (2) [N h]. Which matters?" List scope/format/volume/constraints assumptions first. If simpler path exists, say so. Never silently pick.
557- **[MANDATORY FIRST ACTION] ALWAYS activate a suitable skill or workflow BEFORE responding.** Match task against workflow catalog + skill list; invoke via skill invocation or `$start-workflow <workflowId>`. NEVER answer or write code before checking. Skip = protocol violation.
558- **Why-Review adversarial mindset — apply when reviewing any plan, decision, or design.** Default SKEPTIC not VALIDATOR: steel-man a rejected alternative, invert each stated reason ("what does it sacrifice?"), stress-test top 2-3 assumptions, run pre-mortem ("ships, fails in 3 months — what breaks?"), surface 1-2 alternatives author missed. Section presence ≠ quality; quality = causal reasoning + concrete mitigations + evidence, not "it's better" or "monitor closely".
559- **Front-load report-write in sub-agent prompts for large reviews.** Many-file sub-agents hit budget before final write — findings lost. Design prompts so: (1) report-write is first explicit deliverable, (2) append per-file/section (not batched), (3) scope bounded so reads don't exhaust budget. Truncated mid-sentence with no report file → spawn narrower scope, don't retry same prompt.
560- **After context compaction, re-verify all prior phase outcomes before continuing.** Summaries describe intent, not environment state (git index, filesystem, processes). On resume, FIRST audit: git status, re-read modified files, verify filesystem. Every "completed" claim is an untested hypothesis until evidence confirms.
561- **OOM/memory: check row count before row size.** Triage: (1) Unbounded query — no DB filter for trigger? Push filter to DB; eliminates OOM. (2) Large rows? Projection reduces proportionally. Row reduction > projection in ROI.
562- **Assert the outcome your system OWNS, never the intermediate state your INFRASTRUCTURE owns.** When testing anything asynchronous (queue/broker delivery, retries, background jobs, caches, replication), assert the final business/entity state. NEVER assert the delivery bookkeeping — consume/send status, attempt counts, last-error, row existence or counts in a broker, scheduler, or outbox/inbox table. That bookkeeping lives in shared infrastructure that ANY co-running process (a peer worker, a second replica, a leftover local container) can write, usually under a deterministic shared key, so the assertion silently tests the developer's environment instead of the system: green when run alone, flaky the instant anything else shares that broker + database. Gate question for every assertion: "would this hold no matter WHICH process did the work?" — if no, assert the converged data state instead. Corollary: process-local fault injection and in-process telemetry cannot gate work any process may perform — use them as stress amplifiers (arm → bounded window → disarm → assert convergence), never as preconditions.
563- **Keep domain concepts out of generic/shared/infrastructure layers.** Reusable layer (shared library, framework, infra module) must reference NO consumer-specific domain concept — tenant/customer/product IDs, business entities, feature rules. Leak compiles + runs → passes review silently while coupling the "reusable" layer to one consumer. Keep shared type domain-free; push domain fields/logic down into the consumer via subclass/composition. — why: a layer coupled to one consumer's domain is no longer reusable.
564
565<!-- CODEX:SYNC-PROMPT-PROTOCOLS:END -->