Codex compatibility note:
- Invoke repository skills with
$skill-name in Codex; this mirrored copy rewrites legacy Claude /skill-name references.
- Task tracker mandate: BEFORE executing any workflow or skill step, create/update task tracking for all steps and keep it synchronized as progress changes.
- User-question prompts mean to ask the user directly in Codex.
- Ignore Claude-specific mode-switch instructions when they appear.
- Strict execution contract: when a user explicitly invokes a skill, execute that skill protocol as written.
- Subagent authorization: when a skill is user-invoked or AI-detected and its protocol requires subagents, that skill activation authorizes use of the required
spawn_agent subagent(s) for that task.
- Do not skip, reorder, or merge protocol steps unless the user explicitly approves the deviation first.
- For workflow skills, execute each listed child-skill step explicitly and report step-by-step evidence.
- If a required step/tool cannot run in this environment, stop and ask the user before adapting.
Codex Project-Reference Loading (No Hooks)
Codex uses static project-reference loading instead of runtime-injected project docs.
When coding, planning, debugging, testing, or reviewing, open project docs explicitly using this routing.
Always read:
docs/project-config.json (project-specific paths, commands, modules, and workflow/test settings)
docs/project-reference/docs-index-reference.md (routes to the full docs/project-reference/* catalog)
docs/project-reference/lessons.md (always-on guardrails and anti-patterns)
Missing/stale context route: If docs/project-config.json, the docs index, lessons.md, CLAUDE.md, AGENTS.md, or any task-required reference doc is missing or stale, auto-run $project-init or the narrow setup route ($project-config, $docs-init, $scan-all, $scan --target=<key>, $claude-md-init) before ordinary project-specific work. If Codex mirrors or AGENTS.md are missing/stale, ask the user to run $sync-codex; do not auto-run it.
Situation-based docs:
- Project structure/architecture/tech-stack/deployment/setup (any layer — backend, frontend, or infra):
project-structure-reference.md
- Backend/CQRS/API/domain/entity changes:
backend-patterns-reference.md, domain-entities-reference.md
- Frontend/UI/styling/design-system:
frontend-patterns-reference.md, scss-styling-guide.md, design-system/README.md
- Spec authoring,
docs/specs/ pathing, or TC format: feature-spec-reference.md, spec-system-reference.md, spec-principles.md
- Behavior/public-contract changes or spec-test-code sync:
workflow-spec-test-code-cycle-reference.md plus the spec docs above
- Derived spec indexes/ERDs/reimplementation guides:
spec-system-reference.md and source Feature Specs under docs/specs/
- Integration test implementation/review:
integration-test-reference.md
- E2E test implementation/review:
e2e-test-reference.md
- Code review/audit work:
code-review-rules.md plus domain docs above based on changed files
Do not read all docs blindly. Start from docs-index-reference.md, then open only relevant files for the task.
Quick Summary
Goal: Help users install, configure, troubleshoot, and extend Claude Code CLI (hooks, MCP, skills, commands).
Workflow:
- Categorize — Identify problem type (Setup, Hooks, MCP, Context, Extensibility, Config)
- Diagnose — Follow category-specific diagnostic steps
- Fix & Verify — Apply solution and confirm it resolves the issue
Key Rules:
- Not for writing application code -- use the feature workflow, fix, or refactoring skills instead
- Obtain explicit user approval before modifying settings; never change them unilaterally
- For hooks: check event type, script executability, and JSON output format
Be skeptical. Apply critical thinking, sequential thinking. Every claim needs traced proof, confidence percentages (Idea should be more than 80%).
Claude Code
Purpose
Help users install, configure, troubleshoot, and extend Claude Code CLI -- Anthropic's agentic coding tool with skills, hooks, MCP servers, and slash commands.
When to Use
- Setting up Claude Code for the first time (installation, authentication)
- Troubleshooting hooks that don't fire or produce errors
- Diagnosing MCP server connection failures
- Understanding or resolving context window limits
- Creating or modifying slash commands and agent skills
- Configuring settings (model, allowed tools, output style)
When NOT to Use
- Writing application code -- use the
workflow-feature workflow, fix, or refactoring skills
- Creating or managing MCP servers -- use MCP documentation and project MCP config directly
- AI prompt engineering -- use general prompt-engineering guidance directly
Prerequisites
- Access to
.claude/ directory in the project root
- For hooks: read
.claude/hooks/ directory structure
- For skills: read
.claude/skills/ directory structure
Workflow
Step 1: Identify the Problem Category
| User Says |
Category |
Go To |
| "install", "set up", "authenticate" |
Setup |
Step 2A |
| "hook not firing", "hook error" |
Hook Issues |
Step 2B |
| "MCP not connecting", "MCP error" |
MCP Issues |
Step 2C |
| "context too long", "compaction", "token limit" |
Context Issues |
Step 2D |
| "create skill", "create command" |
Extensibility |
Step 2E |
| "configure", "settings", "model" |
Configuration |
Step 2F |
Step 2A: Setup
- Check prerequisites: Node.js 18+, npm
- Verify authentication:
claude auth status
- IF auth fails: guide through
claude auth login
- Verify project detection: check for
CLAUDE.md in project root
Step 2B: Hook Issues
- Read the hook file causing issues
- Check hook event type matches expected trigger (PreToolUse, PostToolUse, SessionStart, Stop, SubagentStop)
- Verify hook script is executable and has correct shebang
- Check
.claude/settings.json for hook registration
- Test hook in isolation: run the script directly with mock input
- Check for syntax errors in hook output (must be valid JSON for PreToolUse/PostToolUse)
Common fixes:
- Hook not firing: wrong event name or tool matcher pattern
- Hook errors: script not finding dependencies (check relative paths)
- Hook blocks unexpectedly: PreToolUse returning
{ "decision": "block" } incorrectly
Step 2C: MCP Issues
- Check
.claude/settings.json for MCP server configuration
- Verify the MCP server process can start: run the command manually
- Check environment variables (API keys, tokens) are set
- Test connectivity:
claude mcp list to see registered servers
- IF timeout: increase timeout in config or check network
Common fixes:
- "Connection refused": MCP server not running or wrong port
- "Authentication failed": expired or missing API token
- "Tool not found": MCP server registered but tool name mismatch
Step 2D: Context Issues
- Check current context usage (Claude will report when near limit)
- IF approaching limit: suggest
$compact command
- Review if large files are being read unnecessarily
- After compaction, re-read
CLAUDE.md (re-injected each prompt) to re-anchor protocol + workflow catalog
- Run
$recover to resume from on-disk state (active plan, todos) — recovery is skill-driven, not a hook
Step 2E: Extensibility
- For skills: read
references/agent-skills.md for structure
- For custom slash commands: create skills in
.claude/skills/{name}/SKILL.md
- Verify SKILL.md frontmatter has required fields (name, version, description)
- Test: invoke the skill/command and verify it loads
Step 2F: Configuration
- Read
references/configuration.md for settings hierarchy
- Settings locations:
.claude/settings.json (project), ~/.claude/settings.json (user)
- IMPORTANT: Obtain explicit user approval before modifying settings; never change them unilaterally
- Common settings: model selection, allowed tools, output verbosity
Step 3: Verification
- Confirm the fix resolves the user's issue
- Document any configuration changes made
- Warn if changes affect other team members (project-level settings)
Output Format
## Claude Code: [Issue/Task Summary]
### Problem
[What was wrong or what was requested]
### Solution
[Step-by-step fix or setup instructions]
### Files Changed
[List any config files modified, with before/after]
### Verification
[How to confirm the fix works]
Examples
Example 1: Hook Not Firing
User: "My PreToolUse hook for blocking large file reads isn't triggering"
Diagnosis:
- Read
.claude/settings.json -- hook registered under hooks.PreToolUse
- Check tool matcher:
"matcher": "Read" -- correct
- Run script directly:
node .claude/hooks/block-large-reads.cjs -- works
- Found: Hook command uses
%CLAUDE_PROJECT_DIR% but runs from wrong CWD
Fix: Update hook command to use absolute path or verify %CLAUDE_PROJECT_DIR% resolves correctly. Check that the hook entry in settings uses the correct variable syntax for the OS/shell (Windows vs Unix).
Example 2: Setting Up a New Slash Command
User: "I want a /deploy command that runs our staging deployment"
Steps:
- Create
.claude/skills/deploy/SKILL.md:
Deploy to staging environment.
Run the following steps:
1. Verify all tests pass: `npm test`
2. Build the project: `npm run build`
3. Deploy: `npm run deploy:staging`
4. Report deployment status
- Test: type
/deploy in Claude Code CLI
- Verify: command appears in autocomplete and executes the workflow
Reference Files
Load these for detailed guidance on specific topics:
| Topic |
File |
| Installation |
references/getting-started.md |
| Slash commands |
references/slash-commands.md |
| Skills creation |
references/agent-skills.md |
| MCP servers |
references/mcp-integration.md |
| Hooks system |
references/hooks-comprehensive.md |
| Configuration |
references/configuration.md |
| Troubleshooting |
references/troubleshooting.md |
| Enterprise |
references/enterprise-features.md |
Related Skills
skill-creator -- for creating new agent skills with best practices
[IMPORTANT] Use task tracking to break ALL work into small tasks BEFORE starting — including tasks for each file read. This prevents context loss from long files. For simple tasks, AI MUST ATTENTION ask user whether to skip.
AI Mistake Prevention — Failure modes to avoid on every task:
Re-read files after context changes. Context compaction, resume, or long-running work can make memory stale; verify current files before acting.
Verify generated content against source evidence. AI hallucinates APIs, names, claims, and document facts. Check the relevant source before documenting or referencing.
Check downstream references before deleting or renaming. Removing an artifact can stale docs, generated mirrors, configs, and callers; map references first.
Trace the full impact chain after edits. Changing a definition can miss derived outputs and consumers. Follow the affected chain before declaring done.
Verify ALL affected outputs, not just the first. One green check is not all green checks; validate every output surface the change can affect.
Assume existing values are intentional — ask WHY before changing OR flagging one as a defect. Before changing or reporting a constant, limit, flag, cutoff, wording, or pattern, read nearby context and history, the CALLER's ordering, and 2+ sibling call sites of the same convention. A doc stating WHAT without WHY is missing rationale, not proof of a missing guard.
Surface ambiguity before acting — don't pick silently. Multiple valid interpretations require an explicit question or stated assumption with risk.
Assert the outcome your system owns, not the intermediate state your infrastructure owns. When verifying async work, assert the final business state — never the delivery/retry bookkeeping held in shared infrastructure that any co-running process can write. Such a check passes when run alone and flakes the moment anything else shares that infrastructure.
Keep shared guidance role-relevant. Universal guidance must help every receiving skill or agent; code-specific obligations belong only in code-specific protocols.
Critical Thinking Mindset — Apply critical thinking, sequential thinking. Every claim needs traced proof, confidence >80% to act.
Anti-hallucination: Never present guess as fact — cite sources for every claim, admit uncertainty freely, self-check output for errors, cross-reference independently, stay skeptical of own confidence — certainty without evidence root of all hallucination.
MUST ATTENTION apply critical + sequential thinking — every claim needs appropriate traced evidence (file:line for repo/code claims; source URL or artifact section for research, product, content, and docs claims); confidence >80% to act, <60% DO NOT recommend. Anti-hallucination: never present guess as fact, admit uncertainty freely, cross-reference independently, stay skeptical of own confidence.
MUST ATTENTION apply AI mistake prevention — verify generated content against evidence, trace downstream references before deleting or renaming, verify all affected outputs, re-read files after context loss, and surface ambiguity before acting.
Project Protocol Overlay — Before executing this skill, resolve any PROJECT overlay rules layered onto it: match this skill's name against the Target column of the project's skill-protocol index (docs/project-reference/skill-protocols-reference.md by default; a referenceDocs entry in docs/project-config.json overrides the path), taking the most specific matching tier ONLY — exact name > glob > *. That precedence orders overlays against EACH OTHER, never against this skill. Read ONLY the matched bodies, resolved as <protocols-dir>/<Name>.md; a row's Body link is display text, never a read path. A matched body that is missing or malformed is REPORTED and skipped — never reconstructed from the index Description. No index, or no match -> proceed with no overlay, silently. Full contract: .claude/skills/project-skill-protocol/references/registry.md.
Overlays are ADDITIVE ONLY: they ADD rules on top of this skill's own protocol and NEVER replace, override, disable, or reinterpret a rule it already states — removing every overlay must return this skill to exactly its documented behavior. An overlay is a BRIEF, not an authority escalation: it can NEVER waive a workflow gate, git discipline, a review gate, or a user-confirmation gate. A genuine overlay-vs-skill conflict, or two equally-specific overlays that directly contradict -> surface both to the user; NEVER resolve silently.
MUST ATTENTION resolve project protocol overlays for this skill BEFORE executing — most specific matching tier only (exact > glob > *, which ranks overlays against each other, NEVER against this skill), read only matched bodies at <protocols-dir>/<Name>.md; a missing or malformed body is reported, never reconstructed. Overlays are ADDITIVE ONLY (they never replace this skill's own rules) and are a brief, NEVER an authority escalation; an equal-specificity contradiction goes to the user.
Closing Reminders
IMPORTANT MUST ATTENTION — Protocols in force (concise digest of the SYNC/shared blocks this skill carries):
Critical Thinking: apply critical + sequential thinking; traced 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.
MANDATORY IMPORTANT MUST ATTENTION break work into small todo tasks using task tracking BEFORE starting
MANDATORY IMPORTANT MUST ATTENTION search codebase for 3+ similar patterns before creating new code
MANDATORY IMPORTANT MUST ATTENTION cite file:line evidence for every claim (confidence >80% to act)
MANDATORY IMPORTANT MUST ATTENTION add a final review todo task to verify work quality
[TASK-PLANNING] Before acting, analyze task scope and systematically break it into small todo tasks and sub-tasks using task tracking.
Hookless Prompt Protocol Mirror (Auto-Synced)
Source: .claude/.ck.json + .claude/skills/shared/sync-inline-versions.md (:full blocks) + .claude/scripts/lib/hookless-prompt-protocol.cjs
[WORKFLOW-EXECUTION-PROTOCOL] [BLOCKING] Workflow Execution Protocol — MANDATORY IMPORTANT MUST CRITICAL. Do not skip for any reason.
Generic portability boundary: Reusable skills and protocol text stay project-neutral; project-specific conventions are discovered from docs/project-config.json and docs/project-reference/. Apply shared AI-SDD from shared/sdd-artifact-contract.md. Read docs/project-config.json and docs/project-reference/docs-index-reference.md, then open the project reference docs named there. For spec, test-case, behavior-change, public-contract, or docs/specs/ work, route through the local spec docs named by the docs index: feature-spec-reference.md, spec-system-reference.md, spec-principles.md, and workflow-spec-test-code-cycle-reference.md when specs/tests/code must stay synchronized. If either file or a required reference doc is missing or stale, auto-run $project-init (or the narrow lower-level route such as $project-config, $docs-init, $scan-all, or $scan --target=<key>) before ordinary project-specific work. Any supported AI tool may execute when this shared context and local docs are available.
- DETECT: If the prompt starts with an explicit slash skill/workflow command, execute it directly. Otherwise match the prompt against the workflow catalog and skill list.
- ANALYZE: Choose the best option: execute directly, invoke a skill, activate a standard workflow, or compose a custom step combination.
- AUTO-SELECT: Pick the best option yourself. Do not ask the user to choose between direct execution, skill, standard workflow, or custom workflow.
- ACTIVATE: For a selected workflow, call
$start-workflow <workflowId>; for a selected skill, invoke that skill; for a custom workflow, sequence custom steps directly; for direct execution, proceed with the task.
- CREATE TASKS: task tracking for ALL workflow/skill/custom steps before execution when the selected path has multiple steps.
- PARALLELIZE: Before executing the task list, tag each task
PAR (independent inputs + write set disjoint from every other PAR task) or SEQ (name the blocking dependency), group PAR tasks into waves, declare the wave plan, and spawn each wave's sub-agents in ONE message — all-return barrier per wave, fan-out one level deep unless a sub-agent's own definition authorizes further fan-out. Sequential-by-default is a defect when tasks are independent; do not parallelize shared write targets, output-consuming tasks, trivial single-file work, ordering a skill or workflow explicitly fixes, or user-approval gates.
- EXECUTE: Advance per the Workflow Step Advancement & Parallel Phases rule in your context instructions — model-driven; a sub-agent completion advances a step identically to an inline call; a parallel-phase group is an all-return barrier (advance only after ALL members return, never serialize it)
Shared AI-SDD Protocol Markers
Source: .claude/skills/shared/sync-inline-versions.md
SYNC:ai-sdd-artifact-contract
AI-SDD Artifact Contract — Shared spec-driven development rules stay portable and source-owned.
- Keep reusable AI-SDD principles in
.claude; put repository-specific paths, commands, owners, products, and formats in project config/reference docs.
- Preserve cycle:
spec -> plan -> tasks -> implement -> verify -> update spec/docs.
- Trace every requirement or invariant through decision, task, TC/test, source evidence, and docs/spec update.
- Treat code-to-spec extraction as reference-only until accepted by the canonical spec owner.
- Any supported AI tool may plan, implement, review, or verify with synced context; using multiple tools is optional.
- Update
.claude source first, then sync generated mirrors; do not manually edit .agents, .codex, or AGENTS.md. — why: mirrors are generated artifacts; hand-edits are overwritten on the next sync
- If
docs/project-config.json, root instruction files, or a required project-reference doc is missing or stale, auto-run $project-init or the narrow lower-level route before ordinary project-specific work.
Active reference: shared/sdd-artifact-contract.md in the active skills root.
SYNC:ai-sdd-artifact-contract:reminder
- MANDATORY Apply
shared/sdd-artifact-contract.md; keep reusable AI-SDD in .claude and local rules in project docs.
- MANDATORY Code-to-spec extraction is reference-only until canonical acceptance; any supported AI tool may execute with synced context.
- MANDATORY Update
.claude source before syncing generated mirrors; do not manually edit .agents, .codex, or AGENTS.md.
- MANDATORY Missing or stale project config, root instruction files, or required reference docs route project-specific work through
$project-init or the narrow setup route automatically.
[TASK-PLANNING] [MANDATORY] BEFORE executing any workflow or skill step, create/update task tracking for all planned steps, then keep it synchronized as each step starts/completes.
[LESSON-LEARNED-REMINDER] [BLOCKING] Task Planning & Continuous Improvement — MANDATORY. Do not skip.
Break work into small tasks (task tracking) before starting. Add final task: "Analyze AI mistakes & lessons learned".
Extract lessons — ROOT CAUSE ONLY, not symptom fixes:
- Name the FAILURE MODE (reasoning/assumption failure), not symptom — "assumed API existed without reading source" not "used wrong enum value".
- Generality test: does this failure mode apply to ≥3 contexts/codebases? If not, abstract one level up.
- Write as a universal rule — strip project-specific names/paths/classes. Useful on any codebase.
- Consolidate: multiple mistakes sharing one failure mode → ONE lesson.
- Recurrence gate: "Would this recur in future session WITHOUT this reminder?" — No → skip
$learn.
- Auto-fix gate: "Could
$code-review/$code-simplifier/$security-review/$lint catch this?" — Yes → improve review skill instead.
- BOTH gates pass → ask user to run
$learn.
[CRITICAL-THINKING-MINDSET] Apply critical thinking, sequential thinking. Every claim needs traced proof, confidence >80% to act.
Anti-hallucination principle: Never present guess as fact — cite sources for every claim, admit uncertainty freely, self-check output for errors, cross-reference independently, stay skeptical of own confidence — certainty without evidence root of all hallucination.
AI Attention principle (Primacy-Recency): Put the 3 most critical rules at both top and bottom of long prompts/protocols so instruction adherence survives long context windows.
Goal-driven execution: Define success criteria first, loop until verified, and stop only when observable checks pass.
Tests verify intent: Tests must protect business rules/invariants and fail when the protected intent breaks, not only mirror current behavior.
Common AI Mistake Prevention (System Lessons)
- Re-read files after context compaction. Edit requires prior Read in same context; compaction wipes read state. Re-read before editing.
- Grep for old terms after bulk replacements. AI over-trusts find/replace completeness. Grep full repo after bulk edits for missed refs in docs/configs/catalogs.
- Check downstream references before deleting. Deletions cascade doc/code staleness. Map referencing files before removal.
- After memory loss, check existing state before creating new. Compaction wipes prior-work memory. Query current state to resume — never blindly duplicate.
- Verify AI-generated content against actual code. AI hallucinates APIs, class names, method signatures. Grep to confirm existence before documenting/referencing.
- Trace full dependency chain after edits. Changing a definition misses downstream consumers. Trace the full chain.
- When renaming, grep ALL consumer file types. Some file types silently ignore missing refs (no compile error). Search code, templates, configs, generated files.
- Trace ALL code paths when verifying correctness. Code existing ≠ code executing. Trace early exits, error branches, conditional skips — not just happy path.
- Update docs that embed canonical data when source changes. Docs inlining derived data (workflows, schemas, configs) go stale silently. Update all embedding docs alongside source.
- Verify sub-agent results after context recovery. Background agents may finish while parent compacted — grep-verify output, don't trust assumed completion.
- Cross-check full target list against sub-agent assignments. Parallel sub-agents by category miss boundary items. Reconcile union of assignments against target list before proceeding.
- Sub-agents inherit knowledge only from their agent .md definition — use custom agent types, not built-in Explore. Tool adoption = permission + knowledge + enforcement (numbered workflow step).
- Persist sub-agent findings incrementally, not as a final batch. Long sub-agents hit cutoffs before final write — findings lost. Instruct append-per-section to report file.
- When debugging, ask "whose responsibility?" before fixing. Trace caller (wrong data) vs callee (wrong handling). Fix at responsible layer — never patch symptom site.
- Test failure → record a provisional verdict before trace/edit, then investigate. Use the full five-way taxonomy: SOURCE-WRONG (production violates intent), TEST-WRONG (assertion/setup is stale), TEST-NOT-OPTIMAL (valid but fragile or low-signal test), ENVIRONMENT-BLOCKED (external state prevents a verdict), or AMBIGUOUS (intent/evidence cannot choose safely). Then trace root cause and triangulate against the governing spec (
docs/specs/** if one exists) AND source. NEVER weaken an assertion, add a skip, relax a timeout, or change source merely to force green.
- Grep ALL removed names after extraction/refactoring. Primary file "done" ≠ secondary files clean. Grep entire scope for every removed symbol before declaring complete.
- Assume existing values are intentional — ask WHY before changing OR flagging one as a defect. Pattern-matching as "wrong" skips context. Before changing or reporting any constant/limit/flag/cutoff: read comments, git blame, the CALLER's ordering (the guarantee that makes the value correct usually lives in code running immediately BEFORE the cited line), and 2+ sibling call sites of the same convention. A doc stating WHAT without WHY is missing rationale, not proof of a missing guard — and in a validation pass, an accurate
file:line citation proves the transcription, never the defect.
- Verify ALL affected outputs, not just the first. One build green ≠ all green. Multi-stack changes (backend/frontend/tests/docs) require verifying EVERY output.
- Evaluate fit before copying a nearby pattern. Closest example ≠ matching preconditions — verify the new context shares the same constraints, base classes, scope, lifetime.
- Holistic-first debugging — resist nearest-attention trap. Don't dive into first plausible cause. List EVERY precondition (config, env vars, paths, DB, endpoints, creds, versions, DI, data). Verify each against evidence (grep/query — not reasoning). Ask "what would falsify this?" — if nothing, it's not a hypothesis. Most expensive failure: going deeper in "obvious" layer while bug sits in layer never questioned.
- Surgical changes — apply the diff test (context-aware). Two modes: (1) Bug fix → every line traces to the bug; no restyling; orphan cleanup only for imports YOUR changes made unused. (2) Review/enhancement → implement improvements AND announce as "Enhancement beyond main request: [what]". Never silently scope-creep. Diff test: "Would this line exist if I wasn't asked to do X?" — if no, delete or announce.
- Surface ambiguity before coding — don't pick silently. Multiple valid interpretations → present each with effort: "[Request] could mean (1) [N h], (2) [N h]. Which matters?" List scope/format/volume/constraints assumptions first. If simpler path exists, say so. Never silently pick.
- [MANDATORY FIRST ACTION] ALWAYS activate a suitable skill or workflow BEFORE responding. Match task against workflow catalog + skill list; invoke via skill invocation or
$start-workflow <workflowId>. NEVER answer or write code before checking. Skip = protocol violation.
- Why-Review adversarial mindset — apply when reviewing any plan, decision, or design. Default SKEPTIC not VALIDATOR: steel-man a rejected alternative, invert each stated reason ("what does it sacrifice?"), stress-test top 2-3 assumptions, run pre-mortem ("ships, fails in 3 months — what breaks?"), surface 1-2 alternatives author missed. Section presence ≠ quality; quality = causal reasoning + concrete mitigations + evidence, not "it's better" or "monitor closely".
- Front-load report-write in sub-agent prompts for large reviews. Many-file sub-agents hit budget before final write — findings lost. Design prompts so: (1) report-write is first explicit deliverable, (2) append per-file/section (not batched), (3) scope bounded so reads don't exhaust budget. Truncated mid-sentence with no report file → spawn narrower scope, don't retry same prompt.
- After context compaction, re-verify all prior phase outcomes before continuing. Summaries describe intent, not environment state (git index, filesystem, processes). On resume, FIRST audit: git status, re-read modified files, verify filesystem. Every "completed" claim is an untested hypothesis until evidence confirms.
- OOM/memory: check row count before row size. Triage: (1) Unbounded query — no DB filter for trigger? Push filter to DB; eliminates OOM. (2) Large rows? Projection reduces proportionally. Row reduction > projection in ROI.
- Assert the outcome your system OWNS, never the intermediate state your INFRASTRUCTURE owns. When testing anything asynchronous (queue/broker delivery, retries, background jobs, caches, replication), assert the final business/entity state. NEVER assert the delivery bookkeeping — consume/send status, attempt counts, last-error, row existence or counts in a broker, scheduler, or outbox/inbox table. That bookkeeping lives in shared infrastructure that ANY co-running process (a peer worker, a second replica, a leftover local container) can write, usually under a deterministic shared key, so the assertion silently tests the developer's environment instead of the system: green when run alone, flaky the instant anything else shares that broker + database. Gate question for every assertion: "would this hold no matter WHICH process did the work?" — if no, assert the converged data state instead. Corollary: process-local fault injection and in-process telemetry cannot gate work any process may perform — use them as stress amplifiers (arm → bounded window → disarm → assert convergence), never as preconditions.
- Keep domain concepts out of generic/shared/infrastructure layers. Reusable layer (shared library, framework, infra module) must reference NO consumer-specific domain concept — tenant/customer/product IDs, business entities, feature rules. Leak compiles + runs → passes review silently while coupling the "reusable" layer to one consumer. Keep shared type domain-free; push domain fields/logic down into the consumer via subclass/composition. — why: a layer coupled to one consumer's domain is no longer reusable.
1---
2name: claude-code
3description: '[Utilities] Use when you need Claude Code CLI setup, configuration, troubleshooting, and feature guidance.'
4metadata:
5 source: https://skillmd.com/skills/duc01226/claude-code
6---
7
8> Codex compatibility note:
9>
10> - Invoke repository skills with `$skill-name` in Codex; this mirrored copy rewrites legacy Claude `/skill-name` references.
11> - Task tracker mandate: BEFORE executing any workflow or skill step, create/update task tracking for all steps and keep it synchronized as progress changes.
12> - User-question prompts mean to ask the user directly in Codex.
13> - Ignore Claude-specific mode-switch instructions when they appear.
14> - Strict execution contract: when a user explicitly invokes a skill, execute that skill protocol as written.
15> - Subagent authorization: when a skill is user-invoked or AI-detected and its protocol requires subagents, that skill activation authorizes use of the required `spawn_agent` subagent(s) for that task.
16> - Do not skip, reorder, or merge protocol steps unless the user explicitly approves the deviation first.
17> - For workflow skills, execute each listed child-skill step explicitly and report step-by-step evidence.
18> - If a required step/tool cannot run in this environment, stop and ask the user before adapting.
19
20<!-- CODEX:PROJECT-REFERENCE-LOADING:START -->
21
22## Codex Project-Reference Loading (No Hooks)
23
24Codex uses static project-reference loading instead of runtime-injected project docs.
25When coding, planning, debugging, testing, or reviewing, open project docs explicitly using this routing.
26
27**Always read:**
28
29- `docs/project-config.json` (project-specific paths, commands, modules, and workflow/test settings)
30- `docs/project-reference/docs-index-reference.md` (routes to the full `docs/project-reference/*` catalog)
31- `docs/project-reference/lessons.md` (always-on guardrails and anti-patterns)
32
33**Missing/stale context route:** If `docs/project-config.json`, the docs index, `lessons.md`, `CLAUDE.md`, `AGENTS.md`, or any task-required reference doc is missing or stale, auto-run `$project-init` or the narrow setup route (`$project-config`, `$docs-init`, `$scan-all`, `$scan --target=<key>`, `$claude-md-init`) before ordinary project-specific work. If Codex mirrors or `AGENTS.md` are missing/stale, ask the user to run `$sync-codex`; do not auto-run it.
34
35**Situation-based docs:**
36
37- Project structure/architecture/tech-stack/deployment/setup (any layer — backend, frontend, or infra): `project-structure-reference.md`
38- Backend/CQRS/API/domain/entity changes: `backend-patterns-reference.md`, `domain-entities-reference.md`
39- Frontend/UI/styling/design-system: `frontend-patterns-reference.md`, `scss-styling-guide.md`, `design-system/README.md`
40- Spec authoring, `docs/specs/` pathing, or TC format: `feature-spec-reference.md`, `spec-system-reference.md`, `spec-principles.md`
41- Behavior/public-contract changes or spec-test-code sync: `workflow-spec-test-code-cycle-reference.md` plus the spec docs above
42- Derived spec indexes/ERDs/reimplementation guides: `spec-system-reference.md` and source Feature Specs under `docs/specs/`
43- Integration test implementation/review: `integration-test-reference.md`
44- E2E test implementation/review: `e2e-test-reference.md`
45- Code review/audit work: `code-review-rules.md` plus domain docs above based on changed files
46
47Do not read all docs blindly. Start from `docs-index-reference.md`, then open only relevant files for the task.
48
49<!-- CODEX:PROJECT-REFERENCE-LOADING:END -->
50
51## Quick Summary
52
53**Goal:** Help users install, configure, troubleshoot, and extend Claude Code CLI (hooks, MCP, skills, commands).
54
55**Workflow:**
56
571. **Categorize** — Identify problem type (Setup, Hooks, MCP, Context, Extensibility, Config)
582. **Diagnose** — Follow category-specific diagnostic steps
593. **Fix & Verify** — Apply solution and confirm it resolves the issue
60
61**Key Rules:**
62
63- Not for writing application code -- use the feature workflow, fix, or refactoring skills instead
64- Obtain explicit user approval before modifying settings; never change them unilaterally
65- For hooks: check event type, script executability, and JSON output format
66
67**Be skeptical. Apply critical thinking, sequential thinking. Every claim needs traced proof, confidence percentages (Idea should be more than 80%).**
68
69# Claude Code
70
71## Purpose
72
73Help users install, configure, troubleshoot, and extend Claude Code CLI -- Anthropic's agentic coding tool with skills, hooks, MCP servers, and slash commands.
74
75## When to Use
76
77- Setting up Claude Code for the first time (installation, authentication)
78- Troubleshooting hooks that don't fire or produce errors
79- Diagnosing MCP server connection failures
80- Understanding or resolving context window limits
81- Creating or modifying slash commands and agent skills
82- Configuring settings (model, allowed tools, output style)
83
84## When NOT to Use
85
86- Writing application code -- use the `workflow-feature` workflow, `fix`, or `refactoring` skills
87- Creating or managing MCP servers -- use MCP documentation and project MCP config directly
88- AI prompt engineering -- use general prompt-engineering guidance directly
89
90## Prerequisites
91
92- Access to `.claude/` directory in the project root
93- For hooks: read `.claude/hooks/` directory structure
94- For skills: read `.claude/skills/` directory structure
95
96## Workflow
97
98### Step 1: Identify the Problem Category
99
100| User Says | Category | Go To |
101| ----------------------------------------------- | -------------- | ------- |
102| "install", "set up", "authenticate" | Setup | Step 2A |
103| "hook not firing", "hook error" | Hook Issues | Step 2B |
104| "MCP not connecting", "MCP error" | MCP Issues | Step 2C |
105| "context too long", "compaction", "token limit" | Context Issues | Step 2D |
106| "create skill", "create command" | Extensibility | Step 2E |
107| "configure", "settings", "model" | Configuration | Step 2F |
108
109### Step 2A: Setup
110
1111. Check prerequisites: Node.js 18+, npm
1122. Verify authentication: `claude auth status`
1133. IF auth fails: guide through `claude auth login`
1144. Verify project detection: check for `CLAUDE.md` in project root
115
116### Step 2B: Hook Issues
117
1181. Read the hook file causing issues
1192. Check hook event type matches expected trigger (PreToolUse, PostToolUse, SessionStart, Stop, SubagentStop)
1203. Verify hook script is executable and has correct shebang
1214. Check `.claude/settings.json` for hook registration
1225. Test hook in isolation: run the script directly with mock input
1236. Check for syntax errors in hook output (must be valid JSON for PreToolUse/PostToolUse)
124
125**Common fixes:**
126
127- Hook not firing: wrong event name or tool matcher pattern
128- Hook errors: script not finding dependencies (check relative paths)
129- Hook blocks unexpectedly: PreToolUse returning `{ "decision": "block" }` incorrectly
130
131### Step 2C: MCP Issues
132
1331. Check `.claude/settings.json` for MCP server configuration
1342. Verify the MCP server process can start: run the command manually
1353. Check environment variables (API keys, tokens) are set
1364. Test connectivity: `claude mcp list` to see registered servers
1375. IF timeout: increase timeout in config or check network
138
139**Common fixes:**
140
141- "Connection refused": MCP server not running or wrong port
142- "Authentication failed": expired or missing API token
143- "Tool not found": MCP server registered but tool name mismatch
144
145### Step 2D: Context Issues
146
1471. Check current context usage (Claude will report when near limit)
1482. IF approaching limit: suggest `$compact` command
1493. Review if large files are being read unnecessarily
1504. After compaction, re-read `CLAUDE.md` (re-injected each prompt) to re-anchor protocol + workflow catalog
1515. Run `$recover` to resume from on-disk state (active plan, todos) — recovery is skill-driven, not a hook
152
153### Step 2E: Extensibility
154
1551. For skills: read `references/agent-skills.md` for structure
1562. For custom slash commands: create skills in `.claude/skills/{name}/SKILL.md`
1573. Verify SKILL.md frontmatter has required fields (name, version, description)
1584. Test: invoke the skill/command and verify it loads
159
160### Step 2F: Configuration
161
1621. Read `references/configuration.md` for settings hierarchy
1632. Settings locations: `.claude/settings.json` (project), `~/.claude/settings.json` (user)
1643. IMPORTANT: Obtain explicit user approval before modifying settings; never change them unilaterally
1654. Common settings: model selection, allowed tools, output verbosity
166
167### Step 3: Verification
168
169- Confirm the fix resolves the user's issue
170- Document any configuration changes made
171- Warn if changes affect other team members (project-level settings)
172
173## Output Format
174
175```markdown
176## Claude Code: [Issue/Task Summary]
177
178### Problem
179
180[What was wrong or what was requested]
181
182### Solution
183
184[Step-by-step fix or setup instructions]
185
186### Files Changed
187
188[List any config files modified, with before/after]
189
190### Verification
191
192[How to confirm the fix works]
193```
194
195## Examples
196
197### Example 1: Hook Not Firing
198
199**User**: "My PreToolUse hook for blocking large file reads isn't triggering"
200
201**Diagnosis**:
202
2031. Read `.claude/settings.json` -- hook registered under `hooks.PreToolUse`
2042. Check tool matcher: `"matcher": "Read"` -- correct
2053. Run script directly: `node .claude/hooks/block-large-reads.cjs` -- works
2064. **Found**: Hook command uses `%CLAUDE_PROJECT_DIR%` but runs from wrong CWD
207
208**Fix**: Update hook command to use absolute path or verify `%CLAUDE_PROJECT_DIR%` resolves correctly. Check that the hook entry in settings uses the correct variable syntax for the OS/shell (Windows vs Unix).
209
210### Example 2: Setting Up a New Slash Command
211
212**User**: "I want a /deploy command that runs our staging deployment"
213
214**Steps**:
215
2161. Create `.claude/skills/deploy/SKILL.md`:
217
218```markdown
219Deploy to staging environment.
220
221Run the following steps:
222
2231. Verify all tests pass: `npm test`
2242. Build the project: `npm run build`
2253. Deploy: `npm run deploy:staging`
2264. Report deployment status
227```
228
2292. Test: type `/deploy` in Claude Code CLI
2303. Verify: command appears in autocomplete and executes the workflow
231
232## Reference Files
233
234Load these for detailed guidance on specific topics:
235
236| Topic | File |
237| --------------- | ----------------------------------- |
238| Installation | `references/getting-started.md` |
239| Slash commands | `references/slash-commands.md` |
240| Skills creation | `references/agent-skills.md` |
241| MCP servers | `references/mcp-integration.md` |
242| Hooks system | `references/hooks-comprehensive.md` |
243| Configuration | `references/configuration.md` |
244| Troubleshooting | `references/troubleshooting.md` |
245| Enterprise | `references/enterprise-features.md` |
246
247## Related Skills
248
249- `skill-creator` -- for creating new agent skills with best practices
250
251---
252
253> **[IMPORTANT]** Use task tracking to break ALL work into small tasks BEFORE starting — including tasks for each file read. This prevents context loss from long files. For simple tasks, AI MUST ATTENTION ask user whether to skip.
254
255<!-- SYNC:ai-mistake-prevention -->
256
257> **AI Mistake Prevention** — Failure modes to avoid on every task:
258>
259> **Re-read files after context changes.** Context compaction, resume, or long-running work can make memory stale; verify current files before acting.
260> **Verify generated content against source evidence.** AI hallucinates APIs, names, claims, and document facts. Check the relevant source before documenting or referencing.
261> **Check downstream references before deleting or renaming.** Removing an artifact can stale docs, generated mirrors, configs, and callers; map references first.
262> **Trace the full impact chain after edits.** Changing a definition can miss derived outputs and consumers. Follow the affected chain before declaring done.
263> **Verify ALL affected outputs, not just the first.** One green check is not all green checks; validate every output surface the change can affect.
264> **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.
265> **Surface ambiguity before acting — don't pick silently.** Multiple valid interpretations require an explicit question or stated assumption with risk.
266> **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.
267> **Keep shared guidance role-relevant.** Universal guidance must help every receiving skill or agent; code-specific obligations belong only in code-specific protocols.
268
269<!-- /SYNC:ai-mistake-prevention -->
270
271<!-- SYNC:critical-thinking-mindset -->
272
273> **Critical Thinking Mindset** — Apply critical thinking, sequential thinking. Every claim needs traced proof, confidence >80% to act.
274> **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.
275
276<!-- /SYNC:critical-thinking-mindset -->
277
278<!-- SYNC:critical-thinking-mindset:reminder -->
279
280**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.
281
282<!-- /SYNC:critical-thinking-mindset:reminder -->
283
284<!-- SYNC:ai-mistake-prevention:reminder -->
285
286**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.
287
288<!-- /SYNC:ai-mistake-prevention:reminder -->
289
290<!-- SYNC:project-protocol-overlay -->
291
292> **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`.
293>
294> 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.
295
296<!-- /SYNC:project-protocol-overlay -->
297
298<!-- SYNC:project-protocol-overlay:reminder -->
299
300**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.
301
302<!-- /SYNC:project-protocol-overlay:reminder -->
303
304## Closing Reminders
305
306**IMPORTANT MUST ATTENTION — Protocols in force (concise digest of the SYNC/shared blocks this skill carries):**
307
308- **Critical Thinking:** apply critical + sequential thinking; traced proof, confidence >80% to act.
309- **AI Mistake Prevention:** verify generated content against evidence, trace downstream references, verify all affected outputs, re-read after context loss, surface ambiguity.
310
311- **MANDATORY IMPORTANT MUST ATTENTION** break work into small todo tasks using task tracking BEFORE starting
312- **MANDATORY IMPORTANT MUST ATTENTION** search codebase for 3+ similar patterns before creating new code
313- **MANDATORY IMPORTANT MUST ATTENTION** cite `file:line` evidence for every claim (confidence >80% to act)
314- **MANDATORY IMPORTANT MUST ATTENTION** add a final review todo task to verify work quality
315
316**[TASK-PLANNING]** Before acting, analyze task scope and systematically break it into small todo tasks and sub-tasks using task tracking.
317
318<!-- CODEX:SYNC-PROMPT-PROTOCOLS:START -->
319
320## Hookless Prompt Protocol Mirror (Auto-Synced)
321
322Source: `.claude/.ck.json` + `.claude/skills/shared/sync-inline-versions.md` (`:full` blocks) + `.claude/scripts/lib/hookless-prompt-protocol.cjs`
323
324## [WORKFLOW-EXECUTION-PROTOCOL] [BLOCKING] Workflow Execution Protocol — MANDATORY IMPORTANT MUST CRITICAL. Do not skip for any reason.
325
326**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.
327
3281. **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.
3292. **ANALYZE:** Choose the best option: execute directly, invoke a skill, activate a standard workflow, or compose a custom step combination.
3303. **AUTO-SELECT:** Pick the best option yourself. Do not ask the user to choose between direct execution, skill, standard workflow, or custom workflow.
3314. **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.
3325. **CREATE TASKS:** task tracking for ALL workflow/skill/custom steps before execution when the selected path has multiple steps.
3336. **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.
3347. **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)
335
336## Shared AI-SDD Protocol Markers
337
338Source: `.claude/skills/shared/sync-inline-versions.md`
339
340## SYNC:ai-sdd-artifact-contract
341
342> **AI-SDD Artifact Contract** — Shared spec-driven development rules stay portable and source-owned.
343>
344> 1. Keep reusable AI-SDD principles in `.claude`; put repository-specific paths, commands, owners, products, and formats in project config/reference docs.
345> 2. Preserve cycle: `spec -> plan -> tasks -> implement -> verify -> update spec/docs`.
346> 3. Trace every requirement or invariant through decision, task, TC/test, source evidence, and docs/spec update.
347> 4. Treat code-to-spec extraction as reference-only until accepted by the canonical spec owner.
348> 5. Any supported AI tool may plan, implement, review, or verify with synced context; using multiple tools is optional.
349> 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
350> 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.
351>
352> **Active reference:** `shared/sdd-artifact-contract.md` in the active skills root.
353
354---
355
356## SYNC:ai-sdd-artifact-contract:reminder
357
358- **MANDATORY** Apply `shared/sdd-artifact-contract.md`; keep reusable AI-SDD in `.claude` and local rules in project docs.
359- **MANDATORY** Code-to-spec extraction is reference-only until canonical acceptance; any supported AI tool may execute with synced context.
360- **MANDATORY** Update `.claude` source before syncing generated mirrors; do not manually edit `.agents`, `.codex`, or `AGENTS.md`.
361- **MANDATORY** Missing or stale project config, root instruction files, or required reference docs route project-specific work through `$project-init` or the narrow setup route automatically.
362 **[TASK-PLANNING] [MANDATORY]** BEFORE executing any workflow or skill step, create/update task tracking for all planned steps, then keep it synchronized as each step starts/completes.
363
364## [LESSON-LEARNED-REMINDER] [BLOCKING] Task Planning & Continuous Improvement — MANDATORY. Do not skip.
365
366Break work into small tasks (task tracking) before starting. Add final task: "Analyze AI mistakes & lessons learned".
367
368**Extract lessons — ROOT CAUSE ONLY, not symptom fixes:**
369
3701. Name the FAILURE MODE (reasoning/assumption failure), not symptom — "assumed API existed without reading source" not "used wrong enum value".
3712. Generality test: does this failure mode apply to ≥3 contexts/codebases? If not, abstract one level up.
3723. Write as a universal rule — strip project-specific names/paths/classes. Useful on any codebase.
3734. Consolidate: multiple mistakes sharing one failure mode → ONE lesson.
3745. **Recurrence gate:** "Would this recur in future session WITHOUT this reminder?" — No → skip `$learn`.
3756. **Auto-fix gate:** "Could `$code-review`/`$code-simplifier`/`$security-review`/`$lint` catch this?" — Yes → improve review skill instead.
3767. BOTH gates pass → ask user to run `$learn`.
377 **[CRITICAL-THINKING-MINDSET]** Apply critical thinking, sequential thinking. Every claim needs traced proof, confidence >80% to act.
378 **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.
379 **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.
380 **Goal-driven execution:** Define success criteria first, loop until verified, and stop only when observable checks pass.
381 **Tests verify intent:** Tests must protect business rules/invariants and fail when the protected intent breaks, not only mirror current behavior.
382
383## Common AI Mistake Prevention (System Lessons)
384
385- **Re-read files after context compaction.** Edit requires prior Read in same context; compaction wipes read state. Re-read before editing.
386- **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.
387- **Check downstream references before deleting.** Deletions cascade doc/code staleness. Map referencing files before removal.
388- **After memory loss, check existing state before creating new.** Compaction wipes prior-work memory. Query current state to resume — never blindly duplicate.
389- **Verify AI-generated content against actual code.** AI hallucinates APIs, class names, method signatures. Grep to confirm existence before documenting/referencing.
390- **Trace full dependency chain after edits.** Changing a definition misses downstream consumers. Trace the full chain.
391- **When renaming, grep ALL consumer file types.** Some file types silently ignore missing refs (no compile error). Search code, templates, configs, generated files.
392- **Trace ALL code paths when verifying correctness.** Code existing ≠ code executing. Trace early exits, error branches, conditional skips — not just happy path.
393- **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.
394- **Verify sub-agent results after context recovery.** Background agents may finish while parent compacted — grep-verify output, don't trust assumed completion.
395- **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.
396- **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).
397- **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.
398- **When debugging, ask "whose responsibility?" before fixing.** Trace caller (wrong data) vs callee (wrong handling). Fix at responsible layer — never patch symptom site.
399- **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.
400- **Grep ALL removed names after extraction/refactoring.** Primary file "done" ≠ secondary files clean. Grep entire scope for every removed symbol before declaring complete.
401- **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.
402- **Verify ALL affected outputs, not just the first.** One build green ≠ all green. Multi-stack changes (backend/frontend/tests/docs) require verifying EVERY output.
403- **Evaluate fit before copying a nearby pattern.** Closest example ≠ matching preconditions — verify the new context shares the same constraints, base classes, scope, lifetime.
404- **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.
405- **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.
406- **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.
407- **[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.
408- **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".
409- **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.
410- **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.
411- **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.
412- **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.
413- **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.
414
415<!-- CODEX:SYNC-PROMPT-PROTOCOLS:END -->