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.
[IMPORTANT] MUST ATTENTION use council only for multi-option, hard-to-reverse, high-stakes decisions. NEVER council trivial, factual, reversible, or single-option questions.
[IMPORTANT] MUST ATTENTION spawn 5 advisors in parallel, then 5 fresh peer reviewers in parallel, then chairman synthesis.
[IMPORTANT] MUST ATTENTION require evidence for code/architecture claims: file:line, graph trace, or explicit "insufficient evidence."
LLM Council
Quick Summary
Goal: Adversarial decision support for costly wrong choices. Five advisors analyze independently, five fresh reviewers critique anonymously, chairman produces final verdict + report artifacts.
Workflow: Gate → Frame → 5 parallel advisors → anonymized 5-reviewer peer review → chairman synthesis → paired HTML/MD reports.
Key Rules:
- MUST ATTENTION use cheaper ladder first:
$why-review → $plan-validate → $llm-council.
- MUST ATTENTION graph-trace code/architecture questions when
.code-graph/graph.db exists.
- NEVER let earlier advisor responses bleed into later advisors; parallel spawn required.
- ALWAYS mark verdict degraded if fewer than 5 usable advisor responses return.
- ALWAYS regenerate mirrors with
$sync-codex after editing this skill — NEVER hand-edit .agents/ or .codex/ (they are generated artifacts).
Phase 0: Council Gate
Run before advisors.
| Gate |
Required |
Route if false |
| Multi-option |
>=2 viable paths |
Single-option rationale → $why-review |
| Hard to reverse |
Architecture, stack, pricing, hiring, irreversible refactor |
Reversible choice → answer directly |
| Real stakes |
Wrong call costs >=1 week, money, trust, or strategic position |
Low stakes → answer directly |
| Multi-angle value |
Contrarian/first-principles/upside/outside/execution views change answer |
Factual/single-domain → answer directly |
If any gate fails, state failed gate + lighter route. NEVER council "should I use markdown."
Good prompts: pricing model, positioning angle, pivot, hiring vs automation, architecture bet, high-risk launch, irreversible refactor.
Bad prompts: factual lookup, simple yes/no, content generation, summarization, bugfix, package upgrade, routine refactor, anything decidable with one grep.
Advisor Dimensions
Each advisor = thinking dimension, not persona costume. One strong angle each.
| Advisor |
Think |
| Contrarian |
What fails? What assumption kills plan? |
| First Principles Thinker |
What problem are we solving? Which assumptions need rebuild? |
| Expansionist |
What upside, adjacent opportunity, undervalued path is missing? |
| Outsider |
What confuses someone with no context? What jargon hides value? |
| Executor |
What can happen Monday morning? Fastest validated path? |
Tensions: Contrarian vs Expansionist, First Principles vs Executor, Outsider grounds both.
Step 1: Frame Question
When user says "council this", enrich then frame.
Context discovery budget: <=30 seconds; read 2-3 high-signal files.
Search order: CLAUDE.md / AGENTS.md; docs/project-config.json; docs/project-reference/project-structure-reference.md; matching docs/project-reference/*{domain-entities,backend-patterns,frontend-patterns,code-review-rules}*; docs/specs/; memory/; user-referenced files; plans/reports/council-*; domain data (pricing -> revenue, architecture -> service map, tech stack -> dependencies).
Code/architecture gate: If question references existing code, services, files, or blast radius, run before framing:
python .claude/scripts/code_graph trace <key-file> --direction both --json
Skip graph only when .code-graph/graph.db missing or question is non-code.
Framed question includes: core decision, user context, workspace evidence, stakes, constraints, known unknowns. Keep the framing neutral and opinion-free. Ask exactly one clarifying question only if prompt is too vague.
Step 2: Advisor Round
Spawn all 5 advisors simultaneously. Each gets identity, framed question, evidence rules, output constraints.
You are [Advisor Name] on an LLM Council.
Thinking style: [advisor dimension from table]
Question:
---
[framed question]
---
EVIDENCE RULES:
- Code/architecture claims require `file:line`, graph trace, or "I don't have enough evidence yet."
- If existing code context is needed, run:
python .claude/scripts/code_graph trace <file> --direction both --json
python .claude/scripts/code_graph connections <file> --json
- Cite trace output for blast radius, callers, downstream impact.
- Confidence: 95-100% full trace | 80-94% main paths | 60-79% partial | <60% do not recommend.
- Do NOT speculate. Name missing evidence instead.
Respond from assigned angle. Direct, specific, unbalanced by design. Other advisors cover other angles.
Length: 150-300 words plus one-line confidence declaration. No preamble.
Step 3: Peer Review Round
Collect responses. Randomize/anonymize as Response A-E. Spawn 5 fresh reviewers in parallel.
You are reviewing an LLM Council.
Question:
---
[framed question]
---
Anonymized responses:
**Response A:** [response]
**Response B:** [response]
**Response C:** [response]
**Response D:** [response]
**Response E:** [response]
Answer:
1. Strongest response? Why?
2. Biggest blind spot? What is missing?
3. What did all five miss?
Reference responses by letter. Be specific. Under 200 words.
Step 4: Chairman Synthesis
Chairman receives original question, framed question, de-anonymized advisor responses, peer reviews, anonymization map.
You are Chairman of an LLM Council. Synthesize 5 advisors + peer reviews into final verdict.
Question:
---
[framed question]
---
ADVISOR RESPONSES:
**Contrarian:** [response]
**First Principles Thinker:** [response]
**Expansionist:** [response]
**Outsider:** [response]
**Executor:** [response]
PEER REVIEWS:
[all peer reviews]
Produce exact structure:
## Where the Council Agrees
[Independent convergences; high-confidence signals.]
## Where the Council Clashes
[Genuine disagreements. Present both sides; explain why reasonable advisors disagree.]
## Blind Spots the Council Caught
[Only emerged through peer review.]
## The Recommendation
[Clear direct recommendation. Not "it depends."]
## The One Thing to Do First
[Single concrete next step.]
Be direct. Do not hedge. Give clarity unavailable from one perspective.
Quality loop: If chairman misses required sections, ignores degraded-quality state, or makes unsupported code claims, repair with fresh chairman prompt. Max 3 repair rounds; then escalate with missing evidence.
Step 5: Report Artifacts
Write both files; create plans/reports/ if missing.
plans/reports/council-{YYMMDD-HHMM}-{kebab-slug}.html
plans/reports/council-{YYMMDD-HHMM}-{kebab-slug}.md
{YYMMDD-HHMM} = session datetime. {kebab-slug} = 3-6 word question descriptor. Both share prefix. NEVER write artifacts to workspace root or docs/.
HTML: self-contained inline CSS, question top, prominent chairman verdict, agreement/disagreement visual, collapsed advisor responses, collapsed peer review highlights, footer timestamp/topic, professional briefing style, open after generation.
Markdown transcript: original question, framed question, all advisor responses, all peer reviews, anonymization mapping, chairman synthesis.
Output Format
Council complete.
Report: plans/reports/council-{YYMMDD-HHMM}-{kebab-slug}.html
Transcript: plans/reports/council-{YYMMDD-HHMM}-{kebab-slug}.md
Verdict: [1-2 sentence chairman recommendation]
First action: [single next step]
Workflow Integration
Opt-in escalation hook from host skills. NEVER wire into workflow-bugfix, workflow-refactor, or test-* workflows. Blacklist is enforced at the why-review gate (Step A — workflow context check) before the 8-OR frontmatter gate evaluates.
| Host skill |
Mode |
Default |
Gate |
architecture-design |
Always-offer after ## Next Steps |
Skip |
User chooses |
tech-stack-research |
Always-offer after ## Next Steps |
Skip |
User chooses |
domain-analysis |
Always-offer after ## Next Steps |
Skip |
User chooses |
why-review |
Conditional on active plan/PBI frontmatter |
Escalate when gate fires |
Step A workflow blacklist suppression THEN 8-OR frontmatter gate |
prioritize |
Conditional on ranking output |
Escalate when gate fires |
RICE top-2 within 15%, MoSCoW tie, or stakeholder disagreement |
why-review Gate Schema
Gate fires when ANY field true. Absent fields default no-fire; gate opt-in via frontmatter, never opt-out.
| Field |
Convention |
Fires when |
cross_service_impact |
NONE / PARTIAL / FULL |
value != NONE |
breaking_changes |
bool |
true |
complexity |
low / medium / high / critical |
high, critical, or story_points >= 13 |
new_framework |
bool |
true |
irreversible |
bool |
true |
security_critical |
bool |
true |
performance_critical |
bool |
true |
cost_high |
bool |
true |
Host prompt copy MUST cite cheaper rungs: $why-review, $plan-validate, $llm-council.
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.
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.
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
IMPORTANT MUST ATTENTION use council only for multi-option, hard-to-reverse, high-stakes decisions.
Protocols in force — MUST ATTENTION (concise digest of the SYNC/shared blocks this skill carries):
- Critical Thinking: apply critical + sequential thinking; traced
file:line proof, confidence >80% to act.
- AI Mistake Prevention: verify generated content against evidence, trace downstream references, verify all affected outputs, re-read after context loss, surface ambiguity.
IMPORTANT MUST ATTENTION spawn 5 advisors in parallel, then 5 fresh peer reviewers in parallel, then chairman synthesis.
IMPORTANT MUST ATTENTION require evidence for code/architecture claims: file:line, graph trace, or explicit "insufficient evidence."
IMPORTANT MUST ATTENTION mark verdict degraded if fewer than 5 usable advisor responses return.
IMPORTANT MUST ATTENTION write paired HTML + Markdown artifacts under plans/reports/ and open HTML.
IMPORTANT MUST ATTENTION after editing this skill, run $sync-codex to regenerate mirrors — NEVER hand-edit .agents/ or .codex/ (generated).
Anti-Rationalization:
| Evasion |
Rebuttal |
| "This decision feels important" |
Gate it: multi-option, hard-to-reverse, real stakes, multi-angle value. |
| "One advisor can handle it" |
Council value comes from independent angles + anonymous peer review. |
| "Sequential spawn is simpler" |
Sequential spawn contaminates independence. Parallel spawn required. |
| "Four advisors is close enough" |
Missing angle changes verdict quality. Mark degraded. |
| "Evidence would slow us down" |
Unsupported code/architecture claims are speculation. Use graph/file proof or say insufficient evidence. |
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.
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: llm-council
3description: '[Decision Support] Use when pressure-testing irreversible, high-stakes decisions with adversarial AI advisors.'
4metadata:
5 source: https://skillmd.com/skills/duc01226/llm-council
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> **[IMPORTANT]** MUST ATTENTION use council only for multi-option, hard-to-reverse, high-stakes decisions. NEVER council trivial, factual, reversible, or single-option questions.
52> **[IMPORTANT]** MUST ATTENTION spawn 5 advisors in parallel, then 5 fresh peer reviewers in parallel, then chairman synthesis.
53> **[IMPORTANT]** MUST ATTENTION require evidence for code/architecture claims: `file:line`, graph trace, or explicit "insufficient evidence."
54
55# LLM Council
56
57## Quick Summary
58
59**Goal:** Adversarial decision support for costly wrong choices. Five advisors analyze independently, five fresh reviewers critique anonymously, chairman produces final verdict + report artifacts.
60
61**Workflow:** Gate → Frame → 5 parallel advisors → anonymized 5-reviewer peer review → chairman synthesis → paired HTML/MD reports.
62
63**Key Rules:**
64
65- MUST ATTENTION use cheaper ladder first: `$why-review` → `$plan-validate` → `$llm-council`.
66- MUST ATTENTION graph-trace code/architecture questions when `.code-graph/graph.db` exists.
67- NEVER let earlier advisor responses bleed into later advisors; parallel spawn required.
68- ALWAYS mark verdict degraded if fewer than 5 usable advisor responses return.
69- ALWAYS regenerate mirrors with `$sync-codex` after editing this skill — NEVER hand-edit `.agents/` or `.codex/` (they are generated artifacts).
70
71---
72
73## Phase 0: Council Gate
74
75Run before advisors.
76
77| Gate | Required | Route if false |
78| ----------------- | ------------------------------------------------------------------------ | --------------------------------------- |
79| Multi-option | >=2 viable paths | Single-option rationale → `$why-review` |
80| Hard to reverse | Architecture, stack, pricing, hiring, irreversible refactor | Reversible choice → answer directly |
81| Real stakes | Wrong call costs >=1 week, money, trust, or strategic position | Low stakes → answer directly |
82| Multi-angle value | Contrarian/first-principles/upside/outside/execution views change answer | Factual/single-domain → answer directly |
83
84If any gate fails, state failed gate + lighter route. NEVER council "should I use markdown."
85
86Good prompts: pricing model, positioning angle, pivot, hiring vs automation, architecture bet, high-risk launch, irreversible refactor.
87Bad prompts: factual lookup, simple yes/no, content generation, summarization, bugfix, package upgrade, routine refactor, anything decidable with one grep.
88
89---
90
91## Advisor Dimensions
92
93Each advisor = thinking dimension, not persona costume. One strong angle each.
94
95| Advisor | Think |
96| ------------------------ | --------------------------------------------------------------- |
97| Contrarian | What fails? What assumption kills plan? |
98| First Principles Thinker | What problem are we solving? Which assumptions need rebuild? |
99| Expansionist | What upside, adjacent opportunity, undervalued path is missing? |
100| Outsider | What confuses someone with no context? What jargon hides value? |
101| Executor | What can happen Monday morning? Fastest validated path? |
102
103Tensions: Contrarian vs Expansionist, First Principles vs Executor, Outsider grounds both.
104
105---
106
107## Step 1: Frame Question
108
109When user says "council this", enrich then frame.
110
111**Context discovery budget:** <=30 seconds; read 2-3 high-signal files.
112
113**Search order:** `CLAUDE.md` / `AGENTS.md`; `docs/project-config.json`; `docs/project-reference/project-structure-reference.md`; matching `docs/project-reference/*{domain-entities,backend-patterns,frontend-patterns,code-review-rules}*`; `docs/specs/`; `memory/`; user-referenced files; `plans/reports/council-*`; domain data (pricing -> revenue, architecture -> service map, tech stack -> dependencies).
114
115**Code/architecture gate:** If question references existing code, services, files, or blast radius, run before framing:
116
117```bash
118python .claude/scripts/code_graph trace <key-file> --direction both --json
119```
120
121Skip graph only when `.code-graph/graph.db` missing or question is non-code.
122
123**Framed question includes:** core decision, user context, workspace evidence, stakes, constraints, known unknowns. Keep the framing neutral and opinion-free. Ask exactly one clarifying question only if prompt is too vague.
124
125---
126
127## Step 2: Advisor Round
128
129Spawn all 5 advisors simultaneously. Each gets identity, framed question, evidence rules, output constraints.
130
131```text
132You are [Advisor Name] on an LLM Council.
133Thinking style: [advisor dimension from table]
134
135Question:
136---
137[framed question]
138---
139
140EVIDENCE RULES:
141- Code/architecture claims require `file:line`, graph trace, or "I don't have enough evidence yet."
142- If existing code context is needed, run:
143 python .claude/scripts/code_graph trace <file> --direction both --json
144 python .claude/scripts/code_graph connections <file> --json
145- Cite trace output for blast radius, callers, downstream impact.
146- Confidence: 95-100% full trace | 80-94% main paths | 60-79% partial | <60% do not recommend.
147- Do NOT speculate. Name missing evidence instead.
148
149Respond from assigned angle. Direct, specific, unbalanced by design. Other advisors cover other angles.
150Length: 150-300 words plus one-line confidence declaration. No preamble.
151```
152
153---
154
155## Step 3: Peer Review Round
156
157Collect responses. Randomize/anonymize as Response A-E. Spawn 5 fresh reviewers in parallel.
158
159```text
160You are reviewing an LLM Council.
161
162Question:
163---
164[framed question]
165---
166
167Anonymized responses:
168**Response A:** [response]
169**Response B:** [response]
170**Response C:** [response]
171**Response D:** [response]
172**Response E:** [response]
173
174Answer:
1751. Strongest response? Why?
1762. Biggest blind spot? What is missing?
1773. What did all five miss?
178
179Reference responses by letter. Be specific. Under 200 words.
180```
181
182---
183
184## Step 4: Chairman Synthesis
185
186Chairman receives original question, framed question, de-anonymized advisor responses, peer reviews, anonymization map.
187
188```text
189You are Chairman of an LLM Council. Synthesize 5 advisors + peer reviews into final verdict.
190
191Question:
192---
193[framed question]
194---
195
196ADVISOR RESPONSES:
197**Contrarian:** [response]
198**First Principles Thinker:** [response]
199**Expansionist:** [response]
200**Outsider:** [response]
201**Executor:** [response]
202
203PEER REVIEWS:
204[all peer reviews]
205
206Produce exact structure:
207## Where the Council Agrees
208[Independent convergences; high-confidence signals.]
209## Where the Council Clashes
210[Genuine disagreements. Present both sides; explain why reasonable advisors disagree.]
211## Blind Spots the Council Caught
212[Only emerged through peer review.]
213## The Recommendation
214[Clear direct recommendation. Not "it depends."]
215## The One Thing to Do First
216[Single concrete next step.]
217
218Be direct. Do not hedge. Give clarity unavailable from one perspective.
219```
220
221**Quality loop:** If chairman misses required sections, ignores degraded-quality state, or makes unsupported code claims, repair with fresh chairman prompt. Max 3 repair rounds; then escalate with missing evidence.
222
223---
224
225## Step 5: Report Artifacts
226
227Write both files; create `plans/reports/` if missing.
228
229```text
230plans/reports/council-{YYMMDD-HHMM}-{kebab-slug}.html
231plans/reports/council-{YYMMDD-HHMM}-{kebab-slug}.md
232```
233
234`{YYMMDD-HHMM}` = session datetime. `{kebab-slug}` = 3-6 word question descriptor. Both share prefix. NEVER write artifacts to workspace root or `docs/`.
235
236**HTML:** self-contained inline CSS, question top, prominent chairman verdict, agreement/disagreement visual, collapsed advisor responses, collapsed peer review highlights, footer timestamp/topic, professional briefing style, open after generation.
237
238**Markdown transcript:** original question, framed question, all advisor responses, all peer reviews, anonymization mapping, chairman synthesis.
239
240---
241
242## Output Format
243
244```text
245Council complete.
246
247Report: plans/reports/council-{YYMMDD-HHMM}-{kebab-slug}.html
248Transcript: plans/reports/council-{YYMMDD-HHMM}-{kebab-slug}.md
249
250Verdict: [1-2 sentence chairman recommendation]
251First action: [single next step]
252```
253
254---
255
256## Workflow Integration
257
258Opt-in escalation hook from host skills. NEVER wire into `workflow-bugfix`, `workflow-refactor`, or `test-*` workflows. Blacklist is enforced at the `why-review` gate (Step A — workflow context check) before the 8-OR frontmatter gate evaluates.
259
260| Host skill | Mode | Default | Gate |
261| --------------------- | ------------------------------------------ | ------------------------ | ---------------------------------------------------------------- |
262| `architecture-design` | Always-offer after `## Next Steps` | Skip | User chooses |
263| `tech-stack-research` | Always-offer after `## Next Steps` | Skip | User chooses |
264| `domain-analysis` | Always-offer after `## Next Steps` | Skip | User chooses |
265| `why-review` | Conditional on active plan/PBI frontmatter | Escalate when gate fires | Step A workflow blacklist suppression THEN 8-OR frontmatter gate |
266| `prioritize` | Conditional on ranking output | Escalate when gate fires | RICE top-2 within 15%, MoSCoW tie, or stakeholder disagreement |
267
268### `why-review` Gate Schema
269
270Gate fires when ANY field true. Absent fields default no-fire; gate opt-in via frontmatter, never opt-out.
271
272| Field | Convention | Fires when |
273| ---------------------- | -------------------------------------- | ------------------------------------------- |
274| `cross_service_impact` | `NONE` / `PARTIAL` / `FULL` | value != `NONE` |
275| `breaking_changes` | bool | true |
276| `complexity` | `low` / `medium` / `high` / `critical` | `high`, `critical`, or `story_points >= 13` |
277| `new_framework` | bool | true |
278| `irreversible` | bool | true |
279| `security_critical` | bool | true |
280| `performance_critical` | bool | true |
281| `cost_high` | bool | true |
282
283Host prompt copy MUST cite cheaper rungs: `$why-review`, `$plan-validate`, `$llm-council`.
284
285---
286
287<!-- SYNC:critical-thinking-mindset -->
288
289> **Critical Thinking Mindset** — Apply critical thinking, sequential thinking. Every claim needs traced proof, confidence >80% to act.
290> **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.
291
292<!-- /SYNC:critical-thinking-mindset -->
293
294<!-- SYNC:ai-mistake-prevention -->
295
296> **AI Mistake Prevention** — Failure modes to avoid on every task:
297>
298> **Re-read files after context changes.** Context compaction, resume, or long-running work can make memory stale; verify current files before acting.
299> **Verify generated content against source evidence.** AI hallucinates APIs, names, claims, and document facts. Check the relevant source before documenting or referencing.
300> **Check downstream references before deleting or renaming.** Removing an artifact can stale docs, generated mirrors, configs, and callers; map references first.
301> **Trace the full impact chain after edits.** Changing a definition can miss derived outputs and consumers. Follow the affected chain before declaring done.
302> **Verify ALL affected outputs, not just the first.** One green check is not all green checks; validate every output surface the change can affect.
303> **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.
304> **Surface ambiguity before acting — don't pick silently.** Multiple valid interpretations require an explicit question or stated assumption with risk.
305> **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.
306> **Keep shared guidance role-relevant.** Universal guidance must help every receiving skill or agent; code-specific obligations belong only in code-specific protocols.
307
308<!-- /SYNC:ai-mistake-prevention -->
309
310<!-- SYNC:project-protocol-overlay -->
311
312> **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`.
313>
314> 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.
315
316<!-- /SYNC:project-protocol-overlay -->
317
318<!-- SYNC:project-protocol-overlay:reminder -->
319
320**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.
321
322<!-- /SYNC:project-protocol-overlay:reminder -->
323
324## Closing Reminders
325
326**IMPORTANT MUST ATTENTION** use council only for multi-option, hard-to-reverse, high-stakes decisions.
327
328**Protocols in force — MUST ATTENTION (concise digest of the SYNC/shared blocks this skill carries):**
329
330- **Critical Thinking:** apply critical + sequential thinking; traced `file:line` proof, confidence >80% to act.
331- **AI Mistake Prevention:** verify generated content against evidence, trace downstream references, verify all affected outputs, re-read after context loss, surface ambiguity.
332
333**IMPORTANT MUST ATTENTION** spawn 5 advisors in parallel, then 5 fresh peer reviewers in parallel, then chairman synthesis.
334**IMPORTANT MUST ATTENTION** require evidence for code/architecture claims: `file:line`, graph trace, or explicit "insufficient evidence."
335**IMPORTANT MUST ATTENTION** mark verdict degraded if fewer than 5 usable advisor responses return.
336**IMPORTANT MUST ATTENTION** write paired HTML + Markdown artifacts under `plans/reports/` and open HTML.
337**IMPORTANT MUST ATTENTION** after editing this skill, run `$sync-codex` to regenerate mirrors — NEVER hand-edit `.agents/` or `.codex/` (generated).
338
339**Anti-Rationalization:**
340
341| Evasion | Rebuttal |
342| ------------------------------- | -------------------------------------------------------------------------------------------------------- |
343| "This decision feels important" | Gate it: multi-option, hard-to-reverse, real stakes, multi-angle value. |
344| "One advisor can handle it" | Council value comes from independent angles + anonymous peer review. |
345| "Sequential spawn is simpler" | Sequential spawn contaminates independence. Parallel spawn required. |
346| "Four advisors is close enough" | Missing angle changes verdict quality. Mark degraded. |
347| "Evidence would slow us down" | Unsupported code/architecture claims are speculation. Use graph/file proof or say insufficient evidence. |
348
349<!-- SYNC:critical-thinking-mindset:reminder -->
350
351**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.
352
353<!-- /SYNC:critical-thinking-mindset:reminder -->
354<!-- SYNC:ai-mistake-prevention:reminder -->
355
356**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.
357
358<!-- /SYNC:ai-mistake-prevention:reminder -->
359
360<!-- CODEX:SYNC-PROMPT-PROTOCOLS:START -->
361
362## Hookless Prompt Protocol Mirror (Auto-Synced)
363
364Source: `.claude/.ck.json` + `.claude/skills/shared/sync-inline-versions.md` (`:full` blocks) + `.claude/scripts/lib/hookless-prompt-protocol.cjs`
365
366## [WORKFLOW-EXECUTION-PROTOCOL] [BLOCKING] Workflow Execution Protocol — MANDATORY IMPORTANT MUST CRITICAL. Do not skip for any reason.
367
368**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 `$scan --target=<key>`) before ordinary project-specific work. Any supported AI tool may execute when this shared context and local docs are available.
369
3701. **DETECT:** If the prompt starts with an explicit slash skill/workflow command, execute it directly. Otherwise match the prompt against the workflow catalog and skill list.
3712. **ANALYZE:** Choose the best option: execute directly, invoke a skill, activate a standard workflow, or compose a custom step combination.
3723. **AUTO-SELECT:** Pick the best option yourself. Do not ask the user to choose between direct execution, skill, standard workflow, or custom workflow.
3734. **ACTIVATE:** For a selected workflow, call `$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.
3745. **CREATE TASKS:** task tracking for ALL workflow/skill/custom steps before execution when the selected path has multiple steps.
3756. **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.
3767. **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)
377
378## Shared AI-SDD Protocol Markers
379
380Source: `.claude/skills/shared/sync-inline-versions.md`
381
382## SYNC:ai-sdd-artifact-contract
383
384> **AI-SDD Artifact Contract** — Shared spec-driven development rules stay portable and source-owned.
385>
386> 1. Keep reusable AI-SDD principles in `.claude`; put repository-specific paths, commands, owners, products, and formats in project config/reference docs.
387> 2. Preserve cycle: `spec -> plan -> tasks -> implement -> verify -> update spec/docs`.
388> 3. Trace every requirement or invariant through decision, task, TC/test, source evidence, and docs/spec update.
389> 4. Treat code-to-spec extraction as reference-only until accepted by the canonical spec owner.
390> 5. Any supported AI tool may plan, implement, review, or verify with synced context; using multiple tools is optional.
391> 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
392> 7. If `docs/project-config.json`, root instruction files, or a required project-reference doc is missing or stale, auto-run `$project-init` or the narrow lower-level route before ordinary project-specific work.
393>
394> **Active reference:** `shared/sdd-artifact-contract.md` in the active skills root.
395
396---
397
398## SYNC:ai-sdd-artifact-contract:reminder
399
400- **MANDATORY** Apply `shared/sdd-artifact-contract.md`; keep reusable AI-SDD in `.claude` and local rules in project docs.
401- **MANDATORY** Code-to-spec extraction is reference-only until canonical acceptance; any supported AI tool may execute with synced context.
402- **MANDATORY** Update `.claude` source before syncing generated mirrors; do not manually edit `.agents`, `.codex`, or `AGENTS.md`.
403- **MANDATORY** Missing or stale project config, root instruction files, or required reference docs route project-specific work through `$project-init` or the narrow setup route automatically.
404 **[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.
405
406## [LESSON-LEARNED-REMINDER] [BLOCKING] Task Planning & Continuous Improvement — MANDATORY. Do not skip.
407
408Break work into small tasks (task tracking) before starting. Add final task: "Analyze AI mistakes & lessons learned".
409
410**Extract lessons — ROOT CAUSE ONLY, not symptom fixes:**
411
4121. Name the FAILURE MODE (reasoning/assumption failure), not symptom — "assumed API existed without reading source" not "used wrong enum value".
4132. Generality test: does this failure mode apply to ≥3 contexts/codebases? If not, abstract one level up.
4143. Write as a universal rule — strip project-specific names/paths/classes. Useful on any codebase.
4154. Consolidate: multiple mistakes sharing one failure mode → ONE lesson.
4165. **Recurrence gate:** "Would this recur in future session WITHOUT this reminder?" — No → skip `$learn`.
4176. **Auto-fix gate:** "Could `$code-review`/`$code-simplifier`/`$security-review`/`$lint` catch this?" — Yes → improve review skill instead.
4187. BOTH gates pass → ask user to run `$learn`.
419 **[CRITICAL-THINKING-MINDSET]** Apply critical thinking, sequential thinking. Every claim needs traced proof, confidence >80% to act.
420 **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.
421 **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.
422 **Goal-driven execution:** Define success criteria first, loop until verified, and stop only when observable checks pass.
423 **Tests verify intent:** Tests must protect business rules/invariants and fail when the protected intent breaks, not only mirror current behavior.
424
425## Common AI Mistake Prevention (System Lessons)
426
427- **Re-read files after context compaction.** Edit requires prior Read in same context; compaction wipes read state. Re-read before editing.
428- **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.
429- **Check downstream references before deleting.** Deletions cascade doc/code staleness. Map referencing files before removal.
430- **After memory loss, check existing state before creating new.** Compaction wipes prior-work memory. Query current state to resume — never blindly duplicate.
431- **Verify AI-generated content against actual code.** AI hallucinates APIs, class names, method signatures. Grep to confirm existence before documenting/referencing.
432- **Trace full dependency chain after edits.** Changing a definition misses downstream consumers. Trace the full chain.
433- **When renaming, grep ALL consumer file types.** Some file types silently ignore missing refs (no compile error). Search code, templates, configs, generated files.
434- **Trace ALL code paths when verifying correctness.** Code existing ≠ code executing. Trace early exits, error branches, conditional skips — not just happy path.
435- **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.
436- **Verify sub-agent results after context recovery.** Background agents may finish while parent compacted — grep-verify output, don't trust assumed completion.
437- **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.
438- **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).
439- **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.
440- **When debugging, ask "whose responsibility?" before fixing.** Trace caller (wrong data) vs callee (wrong handling). Fix at responsible layer — never patch symptom site.
441- **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.
442- **Grep ALL removed names after extraction/refactoring.** Primary file "done" ≠ secondary files clean. Grep entire scope for every removed symbol before declaring complete.
443- **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.
444- **Verify ALL affected outputs, not just the first.** One build green ≠ all green. Multi-stack changes (backend/frontend/tests/docs) require verifying EVERY output.
445- **Evaluate fit before copying a nearby pattern.** Closest example ≠ matching preconditions — verify the new context shares the same constraints, base classes, scope, lifetime.
446- **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.
447- **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.
448- **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.
449- **[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.
450- **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".
451- **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.
452- **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.
453- **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.
454- **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.
455- **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.
456
457<!-- CODEX:SYNC-PROMPT-PROTOCOLS:END -->