| name | spec-index |
| description | [General] Use when you need to (re)generate a DERIVED navigation index, cross-capability ERD, or reimplementation guide assembled FROM the canonical Feature Specs under docs/specs/**. Never extracts a separate A-E engineering tree. |
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:
- Backend/CQRS/API/domain/entity changes:
backend-patterns-reference.md, domain-entities-reference.md, project-structure-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.
[BLOCKING] Execute skill steps in declared order. NEVER skip, reorder, or merge steps without explicit user approval.
[BLOCKING] Before each step or sub-skill call, update task tracking: set in_progress when step starts, set completed when step ends.
[BLOCKING] Every completed/skipped step MUST include brief evidence or explicit skip reason.
[BLOCKING] If Task tools are unavailable, create and maintain an equivalent step-by-step plan tracker with the same status transitions.
Quick Summary
Portability: docs/specs/ is the fixed Feature Spec root.
Goal: Generate, on demand, a regenerable single-writer navigation layer (catalog + cross-capability ERD + rebuild guide) assembled FROM the canonical tech-free 8-section Feature Specs under docs/specs/{Bucket}/ — so a bucket can be browsed or replatformed without ever forking a second, hand-maintained source of truth. The Feature Specs are the source of truth; this skill only assembles regenerable aids over them.
Summary:
- This skill is a DERIVED-aid assembler only: it reads §1 Overview, §5 Domain Model Mermaid, and §8 TCs from existing Feature Specs to build
INDEX.md (default), an optional cross-capability {Bucket}.erd.md, and an optional {Bucket}.reimplementation-guide.md — it NEVER authors business content.
- Hard prohibition is the load-bearing rule: never emit the retired A-E engineering tree,
M## dirs, 00-module-registry.md, 01-domain-erd.md, 06-reimplementation-guide.md, or docs/specs/README.md/PRIORITY-INDEX.md. Use the {Bucket}.* filenames instead.
- Step 0 Scope Gate via a direct user question (bucket + mode + artifacts) is BLOCKING and runs before any read; if the target bucket has no
README.*.md specs, STOP and route to $spec — never fabricate a spec to index.
- Every generated file carries the
> DERIVED — regenerate via $spec-index; do NOT hand-edit banner, links each row/entity back to its source spec, and keeps INDEX/ERD prose tech-free — only the reimplementation guide may name a target stack.
Renamed: repurposes the former $spec-discovery skill (v4.0.0 derived-aid rewrite) — $spec-discovery no longer resolves as a slash command; use $spec-index.
[SCOPE] This skill assembles a DERIVED index / ERD / reimplementation guide over the canonical Feature Specs. It MUST NOT emit a per-module A-E engineering bundle (A-domain-model, B-business-rules, C-api-contracts, D-events, E-user-journeys), M## directories, 00-module-registry.md, 01-domain-erd.md, or 06-reimplementation-guide.md — those are not part of the spec model; their content lives in the Feature Spec (thin-index-only contract: output is DERIVED — never emit A-E bundle files). Authority: docs/project-reference/spec-system-reference.md.
Inputs: the canonical 8-section Feature Specs (§1 Overview, §5 Domain Model Mermaid, §8 TCs). Code is the technical source of truth — read it ONLY to resolve cross-spec ERD relationships or a reimplementation build order, never to populate a parallel spec layer.
Modes:
| Mode | Trigger | Input | Output |
|---|
index | default — refresh derived aids | Feature Specs in the target bucket(s) | INDEX.md catalog (+ optional ERD / reimplementation guide), all DERIVED |
audit | explicit request — staleness check | Feature Spec mtimes/git vs derived-artifact age | Stale-list report (which derived aids lag their source specs) |
Workflow: $scout (locate specs) → $spec-index (assemble derived aids) → $review-changes → $watzup
Key Rules:
- [BLOCKING] Output is DERIVED and regenerable — every generated file carries a
> DERIVED — regenerate via $spec-index; do NOT hand-edit banner. It is NEVER a second source of truth.
- [BLOCKING] MUST NOT emit
M## dirs, A-E files, 00-module-registry.md, 01-domain-erd.md, 06-reimplementation-guide.md, docs/specs/README.md, or docs/specs/PRIORITY-INDEX.md (all retired). See Hard Prohibitions.
- §1-7 of a Feature Spec are tech-free; the derived INDEX/ERD inherit that. The reimplementation guide is the sole artifact allowed to name a target stack (it is a rebuild guide —
spec-principles.md §3 exception).
- Every catalog row / ERD entity links back to the source Feature Spec; mark
[UNVERIFIED] rather than guessing.
- Read
docs/project-reference/spec-principles.md §3 (tech-agnostic + banned-token list) before writing any prose.
App Bucket Mapping
Derived aids are organized by App Bucket (matches the single-home spec tree). Resolve service→bucket assignments from the canonical table in docs/project-reference/spec-system-reference.md → App Bucket Mapping — do not inline project-specific bucket names in this skill.
Step 0 — Scope Gate (MANDATORY FIRST)
Before reading anything, use a direct user question. Confirm:
| Dimension | Question | Auto-Default |
|---|
| Bucket ★ | Which App Bucket(s) — one bucket, several, or all of docs/specs/? | — must confirm |
| Mode ★ | index (regenerate derived aids) OR audit (report which derived aids are stale vs their source specs)? | index |
| Artifacts | Which derived aids: bucket INDEX.md / cross-capability ERD / reimplementation guide? | INDEX.md only |
| Stack note | (reimplementation guide only) Name a target rebuild stack, or keep stack-neutral build order? | Stack-neutral |
[BLOCKING] If the target bucket has no Feature Specs matching docs/specs/{Bucket}/README.*.md, STOP and route the user to $spec — there is nothing to derive from. NEVER fabricate a spec to index.
Step 1 — Read the Source Feature Specs
Glob docs/specs/{Bucket}/README.*.md → enumerate the canonical specs.- For each spec, read and extract ONLY:
- Capability name + file link
- Summary — first sentence of
## 1. Overview
- Feature code + TC count + status mix from
## 8. Test Specifications
- Entities + relationships from the
## 5. Domain Model ```mermaid block (for ERD assembly)
- Do NOT re-derive business rules, API contracts, or events into new files — those live in the Feature Spec (§1-7) and in code. You are indexing, not extracting.
Scale note: For a bucket with many specs, you MAY spawn parallel reader sub-agents (one per spec) that each return the extracted fields above. This is an optimization, not a gate — there is no per-module A-E extraction to parallelize anymore.
Step 2 — Assemble the Derived Aids
2a. Bucket INDEX.md (default)
Regenerate docs/specs/{Bucket}/INDEX.md as a feature catalog:
> **DERIVED — regenerate via `$spec-index`; do NOT hand-edit.** Source of truth: the Feature Specs in `docs/specs/{Bucket}/README.*.md`.
# {Bucket} — Feature Index
| Capability | Summary | Feature Code | TCs | Status |
| -------------------------- | ------------------- | ------------ | --- | -------------- |
| [{Name}](README.{Name}.md) | {one-line overview} | {FC} | {n} | {Active/Draft} |
$spec owns the canonical Feature Specs only. $spec-index owns this derived INDEX.md and regenerates it deterministically from the specs, so there is one writer for the derived navigation file.
2b. Cross-Capability ERD (on request)
Assemble one Mermaid erDiagram from every spec's §5 block in the bucket:
- Merge entities; dedupe by name; keep cross-capability relationships.
- Resolve a relationship only present implicitly in code by reading the code — but the ERD stays tech-free (entity + relationship names only, no class/table identifiers in prose).
- Write to
docs/specs/{Bucket}/{Bucket}.erd.md with the DERIVED banner. Do NOT name it 01-domain-erd.md (retired).
2c. Reimplementation Guide (on explicit request only)
A build-order narrative: capability dependency order, integration touchpoints, suggested rebuild sequence.
- This is the only derived artifact permitted to name a target stack (rebuild-guide exception,
spec-principles.md §3).
- Write to
docs/specs/{Bucket}/{Bucket}.reimplementation-guide.md with the DERIVED banner. Do NOT name it 06-reimplementation-guide.md (retired).
Step 3 — Stamp & Write
- Every generated file opens with the
> DERIVED — regenerate via $spec-index; do NOT hand-edit banner + a regenerate date.
- Write each file immediately after assembling it; do NOT accumulate large outputs in context.
Step 4 — Verify (self-check before completing)
Hard Prohibitions (R1 mitigation — NON-NEGOTIABLE)
This skill produces only the DERIVED index / ERD / reimplementation guide. Emitting an A-E engineering tree would create a second source of truth competing with the Feature Spec. Therefore this skill MUST NEVER create:
| Forbidden output | Why |
|---|
M## directories (e.g., M01/, M02/) | Retired per-module partition |
A-domain-model.md / B-business-rules.md / C-api-contracts.md / D-events.md / E-user-journeys.md | Retired A-E engineering bundle — content lives in the Feature Spec |
00-module-registry.md | Retired registry — bucket INDEX.md is the catalog |
01-domain-erd.md | Retired per-system ERD name — use {Bucket}.erd.md |
06-reimplementation-guide.md | Retired per-system name — use {Bucket}.reimplementation-guide.md |
docs/specs/README.md, docs/specs/PRIORITY-INDEX.md | Retired QA dashboards — Section 8 is the canonical TC registry |
docs/business-features/** | Not a spec home — all specs live under docs/specs/ |
If a user explicitly asks for an A-E bundle, explain it is retired and offer the derived index/ERD instead. The thin-index-only contract applies — output is DERIVED, never an A-E bundle.
Selective Artifact Mode
| User goal | Generate |
|---|
| "Refresh the bucket index" | INDEX.md only |
| "I need the data model across the app" | Cross-capability ERD ({Bucket}.erd.md) |
| "Produce a rebuild guide" | Reimplementation guide (stack-neutral unless named) |
| "Full navigation set" | INDEX.md + ERD + reimplementation guide |
Next Steps
[BLOCKING] After completing, use a direct user question — DO NOT skip:
- "$docs-update (Recommended)" — sync the Feature Specs + Section 8 if any source content was found stale during indexing
- "$watzup" — wrap up if index generation is the final step
- "Skip, continue manually" — user decides
Related Skills
| Skill | Relationship | When to Call |
|---|
$spec | Source owner — authors the canonical 8-section Feature Spec and flags when derived refresh may be required | Before spec-index — the specs must exist to index |
$spec [mode=tests] | Source owner — owns Section 8 TCs that this skill counts in the catalog | When TCs change and the index TC counts must refresh |
$docs-update | Orchestrator — may call spec-index to refresh derived aids after a doc sync | After code/spec changes need a full doc sync |
$review-changes | Trigger — detects spec changes and surfaces stale derived aids | After spec changes; it will suggest regenerating the index |
What Is Spec Discovery? (v4.0.0)
A derived-index generator over the single-home spec tree. The canonical knowledge is the tech-free 8-section Feature Spec; this skill assembles regenerable navigation aids (catalog, cross-capability ERD, rebuild guide) so readers can browse a bucket or plan a replatform without a second hand-maintained layer. It does NOT reverse-engineer code into a parallel spec bundle — that role was retired with the A-E tree.
Cross-Service Check — Microservices/event-driven: MANDATORY before concluding investigation, plan, spec, or feature doc. Missing downstream consumer = silent regression.
| Boundary | Grep terms |
|---|
| Event producers | Publish, Dispatch, Send, emit, EventBus, outbox, IntegrationEvent |
| Event consumers | Consumer, EventHandler, Subscribe, @EventListener, inbox |
| Sagas/orchestration | Saga, ProcessManager, Choreography, Workflow, Orchestrator |
| Sync service calls | HTTP/gRPC calls to/from other services |
| Shared contracts | OpenAPI spec, proto, shared DTO — flag breaking changes |
| Data ownership | Other service reads/writes same table/collection → Shared-DB anti-pattern |
Per touchpoint: owner service · message name · consumers · risk (NONE / ADDITIVE / BREAKING).
BLOCKED until: Producers scanned · Consumers scanned · Sagas checked · Contracts reviewed · Breaking-change risk flagged
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. Before changing a constant, limit, flag, wording, or pattern, read nearby context and history.
Surface ambiguity before acting — don't pick silently. Multiple valid interpretations require an explicit question or stated assumption with risk.
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.
Prompt-Enhance Closing Anchors
IMPORTANT MUST ATTENTION follow declared step order for this skill; NEVER skip, reorder, or merge steps without explicit user approval
IMPORTANT MUST ATTENTION for every step/sub-skill call: set in_progress before execution, set completed after execution
IMPORTANT MUST ATTENTION every skipped step MUST include explicit reason; every completed step MUST include concise evidence
IMPORTANT MUST ATTENTION if Task tools unavailable, maintain an equivalent step-by-step plan tracker with synchronized statuses
Closing Reminders
- IMPORTANT MUST ATTENTION Goal: Give readers a regenerable, single-writer navigation layer (catalog + cross-capability ERD + rebuild guide) assembled FROM the canonical Feature Specs — so a bucket can be browsed or replatformed without ever forking a second, hand-maintained source of truth
Protocols in force (concise digest of the SYNC/shared blocks this skill carries — MUST ATTENTION each canonical body above):
-
Cross-Service Check: scan producers/consumers/sagas/contracts; flag breaking-change risk.
-
AI Mistake Prevention: verify generated content against evidence, trace downstream references, verify all affected outputs, re-read after context loss, surface ambiguity.
-
Critical Thinking: NEVER present a guess as fact; traced proof, confidence >80% to act.
-
IMPORTANT MUST ATTENTION [BLOCKING] Output is DERIVED — never emit M##/A-E/00-module-registry/01-domain-erd/06-reimplementation-guide/QA-dashboard files (see Hard Prohibitions); use {Bucket}.* filenames instead — why: an A-E bundle becomes a second source of truth competing with the Feature Spec
-
IMPORTANT MUST ATTENTION [BLOCKING] The Feature Spec (docs/specs/{Bucket}/README.{Feature}.md) is the source of truth — this skill assembles, never authors, business content — why: a derived aid that asserts canonical authority corrupts the single-writer contract
-
IMPORTANT MUST ATTENTION [BLOCKING] Confirm bucket + mode + artifacts via a direct user question BEFORE Step 1 — NEVER auto-start; if the bucket has no README.*.md specs, STOP and route to $spec instead of fabricating a spec to index
-
IMPORTANT MUST ATTENTION [BLOCKING] Context compaction/session resume → the current task list FIRST; resume existing tasks, never re-run a completed generation pass — why: summaries describe intent, not filesystem state
-
IMPORTANT MUST ATTENTION [BLOCKING] Stamp the DERIVED banner + regenerate date on every generated file; write after each artifact, never accumulate large outputs in context
-
IMPORTANT MUST ATTENTION [REQUIRED] INDEX/ERD prose tech-agnostic (read spec-principles.md §3 banned-token list FIRST); only a reimplementation guide may name a target stack (rebuild-guide exception)
-
IMPORTANT MUST ATTENTION [REQUIRED] Every catalog row / ERD entity links to an existing Feature Spec — grep the source path to confirm it resolves; mark [UNVERIFIED] rather than guessing — why: dangling links silently rot the navigation layer
-
IMPORTANT MUST ATTENTION Read code ONLY to resolve a cross-spec ERD relationship or a reimplementation build order — never to populate a parallel spec layer — why: code is the technical source of truth, not a spec substitute
Anti-Rationalization:
| Evasion | Rebuttal |
|---|
| "User wants full detail — regenerate the A-E bundle" | A-E is not part of the spec model. Offer the derived index/ERD; the detail already lives in the Feature Spec + code. |
"I'll write the ERD as 01-domain-erd.md" | Wrong filename. Use {Bucket}.erd.md with the DERIVED banner. |
| "The index can be the source of truth, it's complete" | NEVER — it is derived/regenerable. §1-7 + §8 of the Feature Spec are canonical. |
| "No specs in this bucket, I'll extract from code" | STOP. Route to $spec. This skill indexes existing specs; it does not author new ones. |
| "Scope is obvious, skip Step 0 ask the user directly" | BLOCKING — NEVER auto-start. Bucket, mode, and artifact set MUST be confirmed first. |
| "I'll just trust the spec link, no need to verify" | Grep the source path. A dangling link makes the derived navigation layer worse than none. |
| "TC count looks about right, skip re-reading §8" | NEVER guess counts. Re-read ## 8. Test Specifications; mark [UNVERIFIED] if unresolved. |
[TASK-PLANNING] MUST ATTENTION analyze task scope and break into small todo tasks/sub-tasks via task tracking before acting.
[IMPORTANT] Break into many small todo tasks systematically before starting — this is critical.
IMPORTANT MUST ATTENTION Output is DERIVED + regenerable — never a second source of truth; the Feature Spec is canonical.
IMPORTANT MUST ATTENTION Confirm bucket + mode + artifacts via a direct user question BEFORE any read; no specs → STOP, route to $spec.
IMPORTANT MUST ATTENTION Cite file:line for every extracted field; mark [UNVERIFIED] rather than guessing; never emit retired A-E/M##/QA-dashboard files.
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.
- 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.
- 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. Pattern-matching as "wrong" skips context. Before changing any constant/limit/flag: read comments, git blame, surrounding code.
