refactor(claude-md): resolve contradictions, fix dead refs, restructure sections

Fable 5 audit of the global CLAUDE.md (symlinked from this repo):

Contradictions resolved:
- two conflicting graphify sections merged into one (query-first when
  graphify-out/graph.json exists, direct read otherwise; single command
  form; dropped the false 'this project has a knowledge graph' claim)
- plan rule: 'when in doubt skip plan' no longer cancels the mandate —
  borderline = single-file small obvious change
- deviation rule disambiguated: minor/justified -> explain after,
  significant/shaky -> ask before
- 'append-only' registries reconciled with /prune-memory curation

Dead refs fixed: /caveman:compress -> /caveman-compress; design-gate
path now ~/.claude/lib/ (was repo-relative); '(replaces LESSONS)' note
dropped.

Structure: Health Stack / Skill routing / graphify no longer nested
under '# Communication mode'; new '# Tooling & skills' and
'# This repo only' sections; repo-specific Health Stack labeled as such.

Routing updated: + audit-delta, close, capitalize, prune-memory,
profile, context-restore, geo; explicit gstack-OFF fallback rule.
Mid-task question exception generalized to all skill-mandated gates.

Non-critical sections caveman-compressed; Architecture decisions and
Security kept verbatim (must stay unambiguous). Net -1471 chars while
adding 8 routing entries.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

CLAUDE.md

@@ -5,73 +5,104 @@ Apply unless repo-specific instructions override.
 ## Code style
 - Simple, readable, maintainable > clever or compact.
 - One responsibility per function/method.
-- Preserve existing behavior unless asked otherwise.
+- Preserve existing behavior unless asked.
 - Scope changes to task. No unrelated edits.
 
 ## Limits (adapt to language)
-- Max 25 logic lines/function, 80 chars/line, 5 params, 5 local vars (excl. comments, error handling).
+- Max 25 logic lines/function, 80 chars/line, 5 params, 5 local vars.
+  Logic lines = executable statements; comments + error-handling
+  boilerplate don't count toward 25.
 - Too many params → struct/object. Too many vars → split/extract.
 - No global state. Explicit data flow.
 
 ## Comments & readability
-- Document intent, not mechanics. Use project doc style (docstring, JSDoc, etc.).
+- Document intent, not mechanics. Use project doc style (docstring, JSDoc…).
-- Explicit, consistent, meaningful names. Straight control flow. No hidden side effects.
+- Explicit, consistent, meaningful names. Straight control flow.
+  No hidden side effects.
 
 ## Refactoring
 - Priority: safety → readability → consistency.
 - Remove dead code, stale comments, obsolete flags after changes.
-- Non-trivial change: ask "more elegant solution exists?" Hacky fix → rebuild cleanly. No over-engineer.
+- Non-trivial change: ask "more elegant solution exists?"
+  Hacky fix → rebuild clean. No over-engineering.
 
 ## Session start
-1. Read `.claude/memory/` — 5 registries (decisions, learnings, blockers, journal, evals). Apply before touching anything.
+1. Read `.claude/memory/` — 5 registries (decisions, learnings, blockers,
-2. Read `.claude/tasks/TODO.md` — understand current state.
+   journal, evals). Apply before touching anything.
-3. If `.claude/memory/` or `.claude/tasks/TODO.md` missing, create before starting (templates at `~/.claude/templates/memory/`).
+2. Read `.claude/tasks/TODO.md` — current state.
+3. Either missing → create before starting
+   (templates: `~/.claude/templates/memory/`).
 
 ## Workflow
-- Write/modify task touching logic (new behavior, control flow, state, API, dependencies): plan in `.claude/tasks/TODO.md` first, decomposed into subtasks. Task count irrelevant — one complex task still needs plan. When in doubt about complexity, skip plan (be pragmatic).
+- Task touches logic (new behavior, control flow, state, API,
-- Confirm before implementing only when real trade-offs exist (multiple valid approaches, breaking change, destructive action) — else proceed.
+  dependencies) → plan in `.claude/tasks/TODO.md` first, decomposed into
-- Exempt from `.claude/tasks/TODO.md`: pure reads, explanations, questions, typos, cosmetic CSS tweaks, single config-value changes. Aligns with `/hotfix` scope (≤2 files, obvious fix).
+  subtasks. One complex task still needs plan. Borderline case (single
+  file, small obvious logic change) → skip plan, stay pragmatic.
+- Confirm before implementing ONLY when real trade-offs exist (multiple
+  valid approaches, breaking change, destructive action) — else proceed.
+- Exempt from TODO.md: pure reads, explanations, questions, typos,
+  cosmetic CSS, single config-value change. Same scope as `/hotfix`
+  (≤2 files, obvious fix).
 - Minimal changes unless broader refactor requested. State trade-offs.
-- Use sub-agents to keep main context clean — one task per sub-agent. Invest more compute on hard problems.
+- Sub-agents keep main context clean — one task per sub-agent.
-- One question upfront if needed — never interrupt mid-task. *(Exception: orchestrators' mandatory validation gates — example, /init-project STEP 4/7, /ship-feature STEP 3 — exempt.)*
+  More compute on hard problems.
-- Bug received → fix directly: check logs, find root cause, resolve autonomously.
+- One question upfront if needed — never interrupt mid-task.
-- Something goes wrong: STOP and re-plan — never push through.
+  *Exception: skill-mandated gates and checkpoints (orchestrator
-- Report deviations: minor/justified → explain. Significant/unjustified → ask.
+  validation gates, approval gates, darwin checkpoints) always fire.*
-- Root causes only. No temp fixes. Never assume — verify paths, APIs, variables before use.
+- Bug received → fix directly: check logs, find root cause, resolve
+  autonomously.
+- Something goes wrong → STOP, re-plan. Never push through.
+- Deviations: minor or clearly justified → do, explain after.
+  Significant or shaky justification → ask BEFORE deviating.
+- Root causes only. No temp fixes. Never assume — verify paths, APIs,
+  variables before use.
 
 ## After code changes
 1. Run tests, lint, build, type-check if available.
 2. Report what verified, what not.
-3. List remaining risks and surviving deviations.
+3. List remaining risks, surviving deviations.
-4. Never mark complete without proof it works. Bar: "would staff engineer approve this?"
+4. Never mark complete without proof it works.
-5. After any correction or notable event, capitalize to right registry in `.claude/memory/` (see "Memory registries" below).
+   Bar: "would staff engineer approve?"
+5. Correction or notable event → capitalize to right registry
+   (see "Memory registries").
 
 ## Task tracking (`.claude/tasks/TODO.md`)
 
-Applies to any write/modify task touching logic, regardless of count. Skip for reads and trivial edits (see Workflow).
+Any write/modify task touching logic. Skip reads + trivial edits
+(see Workflow).
 
-1. Plan → write task in `.claude/tasks/TODO.md` before touching code.
+1. Plan → task written before code.
-2. Decompose → split each complex task into subtasks (one subtask = one coherent change).
+2. Decompose → one subtask = one coherent change.
-3. Track → check off subtasks as you go.
+3. Track → check off as you go.
-4. Summarize → high-level summary at each milestone.
+4. Summarize → high-level note at each milestone.
 
 ## Memory registries (`.claude/memory/`)
 
-Five append-only registries persist across sessions. Read all at session start. Capitalize during/after work.
+Five registries persist across sessions. Read all at session start.
+Capitalize during/after work. Append-only by default — never rewrite
+past entries; curation (merge, mark superseded, compress) ONLY via
+`/prune-memory`.
 
 | File | ID format | Purpose |
 |------|-----------|---------|
-| `decisions.md` | BDR-XXX | Design/architecture choices with rationale + alternatives + status |
+| `decisions.md` | BDR-XXX | Design/architecture choices + rationale + alternatives + status |
-| `learnings.md` | LRN-XXX | Reusable patterns observed + context + future application (replaces LESSONS) |
+| `learnings.md` | LRN-XXX | Reusable patterns + context + future application |
 | `blockers.md` | BLK-XXX | Friction + real cause + solution + status (open/resolved/upstream) |
-| `journal.md` | date heading | 3-5 lines/session — what done, decided, blocked |
+| `journal.md` | date heading | 3-5 lines/session — done, decided, blocked |
 | `evals.md` | EVAL-XXX | Quality check of Claude's output + method + anomalies + action |
 
-**Language — registries ALWAYS English:**
+**Language — registries ALWAYS English.** Rationale: consistent vocab,
-All persisted entries (decisions, learnings, blockers, journal, evals) must be English. Rationale: consistent vocab for re-read efficiency, lower token cost, easier cross-project reuse. User-facing CAPITALIZE prompts may mirror user's language; only final written entry is English.
+lower token cost, cross-project reuse. User-facing CAPITALIZE prompts may
+mirror user's language; final written entry English.
 
-**Format — registries ALWAYS caveman:**
+**Format — registries ALWAYS caveman.** Drop articles + filler, fragments
-All writes to `.claude/memory/*.md` (decisions, learnings, blockers, journal, evals) MUST use caveman style — drop articles (a/an/the), drop filler (just/really/basically/actually/simply), fragments OK, short synonyms (big not extensive, fix not "implement a solution for"). Keep technical terms exact, code blocks unchanged, error messages quoted exact, IDs (BDR-XXX, LRN-XXX, BLK-XXX, EVAL-XXX) and dates unchanged. Pattern: `[thing] [action] [reason]. [next step].` Rationale: registries loaded every session start — caveman cuts ~40% input tokens with zero loss of technical substance. Applies to direct writes AND skill-driven CAPITALIZE steps (close, ship-feature, feat, bugfix, hotfix, commit-change). Existing entries: compress on demand via `/caveman:compress <file>`.
+OK, short synonyms. Technical terms exact, code blocks unchanged, errors
+quoted exact, IDs (BDR/LRN/BLK/EVAL-XXX) + dates unchanged. Pattern:
+`[thing] [action] [reason]. [next step].` Rationale: registries load
+every session — caveman cuts ~40% input tokens, zero substance loss.
+Applies to direct writes AND skill CAPITALIZE steps (close, ship-feature,
+feat, bugfix, hotfix, commit-change). Existing entries: compress via
+`/caveman-compress <file>`.
 
 **Routing — what goes where:**
 - Choice with tradeoffs you'd defend → `decisions.md`.
@@ -81,47 +112,57 @@ All writes to `.claude/memory/*.md` (decisions, learnings, blockers, journal, ev
 - Did Claude's output actually work? → `evals.md`.
 
 **Proactive capitalization (Claude's responsibility):**
-After any substantive milestone (bug fix with real root cause, feature shipped, non-trivial commit, design choice, surprising discovery, dead end with lesson), **offer to capitalize inline** — do NOT wait for user. Use right registry; pre-fill entry from conversation context; let user approve/edit before writing.
+After substantive milestone (bug fix with real root cause, feature
+shipped, non-trivial commit, design choice, surprising discovery, dead
+end with lesson) → **offer to capitalize inline**, do NOT wait for user.
+Pre-fill entry from context; user approves/edits before write.
+Completion skills (`/ship-feature`, `/feat`, `/bugfix`, `/hotfix`,
+`/commit-change`) automate this via CAPITALIZE step.
 
-Completion skills (`/ship-feature`, `/feat`, `/bugfix`, `/hotfix`, `/commit-change`) include CAPITALIZE step that automates this.
+**Session-close ritual** (`/close`, or inline when asked):
-
+1. What decided? → `decisions.md` (if non-trivial).
-**Session-close ritual** (invoke manually via `/close`, or answer inline when asked):
+2. What learned? → `learnings.md` (if reusable).
-1. What did I decide? → `decisions.md` (if non-trivial).
+3. What blocked? → `blockers.md`.
-2. What did I learn? → `learnings.md` (if reusable).
-3. What am I blocked on? → `blockers.md`.
-
-## Context Navigation (graphify)
-- Use `/graphify query` ONLY for large-scope tasks: multi-file features, complex bug investigations, architectural changes, major refactors.
-- Small tasks (hotfix, typo, single-file change, quick lookup): read files directly — do NOT invoke graphify.
-- Before answering architecture/codebase questions, read `graphify-out/GRAPH_REPORT.md` for god nodes and community structure. If `graphify-out/wiki/index.md` exists, use as navigation entrypoint.
-- After modifying code files, run `/graphify <path> --update` to keep graph current (AST-only, no API cost for code-only changes).
 
 ---
 
 # Architecture decisions
 
-Override default framework/tooling choices. Apply at project creation, scaffolding, brainstorming.
+Override default framework/tooling choices. Apply at project creation,
+scaffolding, brainstorming.
 
 ## Public websites — never SPA
 
-When project is public-facing website meant to be indexed (landing page, portfolio, blog, e-commerce, docs):
+When project is public-facing website meant to be indexed (landing page,
+portfolio, blog, e-commerce, docs):
 
-- **FORBIDDEN**: pure SPA (CRA, Vite React SPA, Vue SPA) for public pages. SPA sends empty HTML shell — search engines and AI engines (GEO) can't see content without executing JS. SEO and AI visibility destroyed.
+- **FORBIDDEN**: pure SPA (CRA, Vite React SPA, Vue SPA) for public pages.
-- **Astro** = default for informational sites (portfolio, docs, blog, landing). Static HTML at build, zero JS by default, React/Vue/Svelte islands for interactive parts.
+  SPA sends empty HTML shell — search engines and AI engines (GEO) can't
-- **Next.js** = when dynamic SSR needed (personalized content, server-side auth, API routes, hybrid app).
+  see content without executing JS. SEO and AI visibility destroyed.
-- **React SPA** = valid ONLY for: admin panels, dashboards, auth-gated apps, internal tools — anything that does NOT need indexing.
+- **Astro** = default for informational sites (portfolio, docs, blog,
-- **Mixed project** (public + admin): Astro/Next for public, React island (`client:only`) for admin.
+  landing). Static HTML at build, zero JS by default, React/Vue/Svelte
-- At brainstorming (`/init-project` STEP 1, `/ship-feature` STEP 1): if project is public website and user hasn't specified framework, PROPOSE Astro and EXPLAIN why not SPA. Never silently pick React CRA.
+  islands for interactive parts.
+- **Next.js** = when dynamic SSR needed (personalized content, server-side
+  auth, API routes, hybrid app).
+- **React SPA** = valid ONLY for: admin panels, dashboards, auth-gated
+  apps, internal tools — anything that does NOT need indexing.
+- **Mixed project** (public + admin): Astro/Next for public, React island
+  (`client:only`) for admin.
+- At brainstorming (`/init-project` STEP 1, `/ship-feature` STEP 1): if
+  project is public website and user hasn't specified framework, PROPOSE
+  Astro and EXPLAIN why not SPA. Never silently pick React CRA.
 
 ## Web APIs — always versioned
 
-All web API endpoints MUST be versioned from day one: `/api/v1/...`, `/api/v2/...`.
+All web API endpoints MUST be versioned from day one: `/api/v1/...`.
 
 - New project → start at `/api/v1/`. No bare `/api/` routes.
-- Breaking changes → new version (`v2`). Old version stays functional — clients migrate at own pace.
+- Breaking changes → new version (`v2`). Old version stays functional —
-- Non-breaking additions (new fields, new endpoints) → keep in current version.
+  clients migrate at own pace.
-- Each version is self-contained contract. Never modify existing version behavior to match newer one.
+- Non-breaking additions (new fields, new endpoints) → current version.
-- Router structure must reflect versioning explicitly (e.g. `api/v1/routes/`, `api/v2/routes/`).
+- Each version is self-contained contract. Never modify existing version
+  behavior to match newer one.
+- Router structure reflects versioning explicitly (e.g. `api/v1/routes/`).
 
 ## Security — non-negotiable defaults
 
@@ -129,26 +170,34 @@ Apply at every dev step: design, scaffolding, implementation, review.
 
 ### Input & data
 - Never trust user input. Validate type, length, format, range before use.
-- Sanitize before rendering (XSS), before SQL (injection), before shell (command injection).
+- Sanitize before rendering (XSS), before SQL (injection), before shell
-- Use parameterized queries / prepared statements. String concatenation into SQL = immediate blocker.
+  (command injection).
+- Use parameterized queries / prepared statements. String concatenation
+  into SQL = immediate blocker.
 
 ### Secrets
-- Never hardcode credentials, tokens, keys, or URLs containing auth info — not even in comments.
+- Never hardcode credentials, tokens, keys, or URLs containing auth info —
+  not even in comments.
 - Always use env vars. Provide `.env.example` with placeholder values only.
 - If secret appears in code during review, flag and stop — do not proceed.
 
 ### Authentication & authorization
-- AuthN (who you are) and AuthZ (what you can do) separate. Never assume AuthN implies AuthZ.
+- AuthN (who you are) and AuthZ (what you can do) separate. Never assume
-- Check authorization on every sensitive endpoint/function — not just at entry point.
+  AuthN implies AuthZ.
+- Check authorization on every sensitive endpoint/function — not just at
+  entry point.
 - Default to deny. Explicit allowlist > implicit denylist.
 
 ### Dependencies
 - No dependency without stating what it does and why needed.
-- Prefer well-maintained, widely-used packages. Flag abandoned or single-maintainer packages.
+- Prefer well-maintained, widely-used packages. Flag abandoned or
-- Never `npm install` or `pip install` package found in random code snippet without naming it explicitly.
+  single-maintainer packages.
+- Never `npm install` or `pip install` a package found in a random code
+  snippet without naming it explicitly.
 
 ### Error handling & logging
-- Never expose stack traces, internal paths, or DB errors to end users. Log internally, return generic message.
+- Never expose stack traces, internal paths, or DB errors to end users.
+  Log internally, return generic message.
 - Never log secrets, passwords, tokens, or PII — even at DEBUG level.
 - Fail closed: on unexpected error, deny access rather than grant.
 
@@ -160,80 +209,109 @@ Apply at every dev step: design, scaffolding, implementation, review.
 
 # Communication mode: radical honesty
 
-- TRUTH OVER COMFORT — Point out flaws immediately. No sugarcoating, no "not bad but…".
+- TRUTH OVER COMFORT — Point out flaws immediately. No sugarcoating,
-- ZERO COMPLACENCY — Never validate idea just because I proposed it. Evaluate arguments on merit.
+  no "not bad but…".
-- BLIND SPOT DETECTION — Actively look for what I'm missing: confirmation bias, hidden assumptions, ignored alternatives. Flag without waiting for permission.
+- ZERO COMPLACENCY — Never validate idea just because I proposed it.
-- ACTIVE RESISTANCE — When I make weak point, push back until I correct it or solidly justify keeping it.
+  Evaluate arguments on merit.
-- UNCERTAINTY TRANSPARENCY — If you don't know, say so. No invention, no vague answers to save face.
+- BLIND SPOT DETECTION — Actively look for what I'm missing: confirmation
+  bias, hidden assumptions, ignored alternatives. Flag without waiting
+  for permission.
+- ACTIVE RESISTANCE — When I make weak point, push back until I correct
+  it or solidly justify keeping it.
+- UNCERTAINTY TRANSPARENCY — If you don't know, say so. No invention,
+  no vague answers to save face.
 
-## Health Stack
+---
 
-- shell: shellcheck *.sh hooks/*.sh lib/*.sh
+# Tooling & skills
 
 ## Skill routing
 
-When user's request matches available skill, ALWAYS invoke via Skill
+Request matches available skill → invoke via Skill tool FIRST. No direct
-tool as FIRST action. Do NOT answer directly, do NOT use other tools first.
+answer, no other tools before. Skill workflows beat ad-hoc answers.
-Skill has specialized workflows that produce better results than ad-hoc answers.
 
 Key routing rules:
-- Product ideas, "is this worth building", brainstorming → invoke office-hours
+- Product ideas, "worth building?", brainstorming → office-hours
-- Bugs, errors, "why is this broken", 500 errors → invoke investigate, or bugfix if no gstack
+- Bugs, errors, "why broken", 500s → investigate, or bugfix if no gstack
-- Small feature (1-5 files, lightweight) → invoke feat
+- Small feature (1-5 files) → feat
-- Quick fix (typo, CSS, config, max 2 files) → invoke hotfix
+- Quick fix (typo, CSS, config, ≤2 files) → hotfix
-- Ship, deploy, push, create PR → invoke ship, or ship-feature if no gstack
+- Ship, deploy, push, PR → ship, or ship-feature if no gstack
-- QA, test site, find bugs → invoke qa
+- QA, test site, find bugs → qa
-- Code review, check my diff → invoke review
+- Code review, check diff → review
-- Update docs after shipping → invoke document-release, or doc if no gstack
+- Docs update post-ship → document-release, or doc if no gstack
-- Stale docs audit, doc sync → invoke doc
+- Stale docs audit, doc sync → doc
-- Weekly retro → invoke retro
+- Recurring audit of changes since last run → audit-delta
-- Design system, brand → invoke design-consultation, then full design toolchain (see "Design work" below)
+- Weekly retro → retro
-- Build UI / component / page / screen → full design toolchain (see "Design work" below)
+- Design system, brand → design-consultation, then design toolchain below
-- Visual audit, design polish → invoke design-review + emil-design-eng lens + design-motion-principles (audit mode)
+- Build UI / component / page / screen → design toolchain below
-- Architecture review → invoke plan-eng-review
+- Visual audit, polish → design-review + emil-design-eng
-- Save progress, checkpoint, resume → invoke context-save
+  + design-motion-principles (audit mode)
-- Code quality, health check → invoke health
+- Architecture review → plan-eng-review
-- Refactor without behavior change → invoke refactor
+- Save/restore working context → context-save / context-restore
-- Dead code, style cleanup → invoke code-clean
+- End-of-session ritual → close
-- SEO/GEO audit → invoke seo
+- Flush memory before /clear or /compact → capitalize
-- Web hardening (SSL/TLS, HSTS, CSP, HTTP→HTTPS, canonical, 404, .htaccess/nginx/vercel/netlify headers+redirects) → invoke harden
+- Registries too big/noisy → prune-memory
-- W3C standards + WCAG a11y (HTML validity, CSS validity, accessibility audit, axe, pa11y, validator.w3.org, normes W3C) → invoke validate
+- Skill profiles (design/dev/qa/minimal) → profile
-- Deep analysis before any modification → invoke analyze
+- Code quality dashboard → health
-- Smart commit grouping → invoke commit-change
+- Refactor without behavior change → refactor
-- Security audit (secrets, deps CVE, OWASP code-level) → invoke cso
+- Dead code, style cleanup → code-clean
-- Initialize new project from scratch → invoke init-project
+- SEO/GEO audit → seo (GEO only → geo)
-- Onboard existing project (config + archetype detection + full audit pipeline + backlog) → invoke onboard
+- Web hardening (SSL/TLS, HSTS, CSP, redirects, headers) → harden
+- W3C + WCAG a11y (HTML/CSS validity, axe, pa11y) → validate
+- Deep analysis before modification → analyze
+- Smart commit grouping → commit-change
+- Security audit (secrets, deps CVE, OWASP) → cso
+- New project from scratch → init-project
+- Onboard existing repo (config + archetype + audits + backlog) → onboard
 
-Design gate (automatic):
+gstack OFF → gstack skills (investigate, ship, qa, review, health, retro,
-All lightweight skills (feat, hotfix, bugfix) include design gate that auto-detects
+office-hours, context-save…) unavailable: use the non-gstack fallback
-UI/style signals in task. If design signals found and ui-ux-pro-max inactive,
+where listed, else say so instead of improvising.
-agent asks user whether to activate before proceeding.
-Gate spec: lib/design-gate.md. Orchestrators (ship-feature, init-project) already
-handle via STEP 0 plugin-check.
 
-Design work — full toolchain (tiered by scope):
+Design gate (automatic): lightweight skills (feat, hotfix, bugfix) detect
-When a task touches design/UI, mobilize the available tools by scope. Reinforced by
+UI/style signals; signals found + ui-ux-pro-max inactive → ask user
-the design-toolchain-reminder UserPromptSubmit hook (injects this on UI signals).
+before proceeding. Gate spec: `~/.claude/lib/design-gate.md`.
-- Trivial (≤2 files, single cosmetic value, CSS tweak) → /hotfix, NO toolchain.
+Orchestrators (ship-feature, init-project) handle via STEP 0 plugin-check.
-- Build UI (new component, page, screen, redesign) → invoke ui-ux-pro-max (plan/build)
+
-  + frontend-design (anti AI-slop) + Magic MCP `/ui` (21st.dev component scaffold)
+## Design work — full toolchain (tiered by scope)
-  + emil-design-eng (polish pass, invisible details) + design-motion-principles
+
-  (when motion/transitions) + design-html (when static HTML / Pretext-native).
+Task touches design/UI → mobilize tools by scope. Reinforced by
-- Design system / brand → design-consultation FIRST (aesthetic, type, color, spacing,
+design-toolchain-reminder hook (injects on UI signals).
-  motion), THEN the build tools above.
+- Trivial (≤2 files, single cosmetic value) → /hotfix, NO toolchain.
+- Build UI (new component, page, screen, redesign) → ui-ux-pro-max
+  (plan/build) + frontend-design (anti AI-slop) + Magic MCP `/ui`
+  (21st.dev scaffold) + emil-design-eng (polish pass)
+  + design-motion-principles (when motion) + design-html (static HTML).
+- Design system / brand → design-consultation FIRST (aesthetic, type,
+  color, spacing, motion), THEN build tools above.
 - Review / audit → design-review (visual QA + fix) + emil-design-eng lens
-  + design-motion-principles (audit mode, catch slop motion).
+  + design-motion-principles (audit mode).
-In doubt about scope (is this a trivial tweak or a real UI change?) → do NOT silently
+Scope doubt (trivial tweak vs real UI change?) → do NOT silently skip
-skip the toolchain. Ask the user, or default to the Build tier rather than /hotfix.
+toolchain: ask user, or default to Build tier.
-Magic MCP (@21st-dev/magic) is configured globally and costs API calls — use it for
+Magic MCP (@21st-dev/magic) costs API calls — component generation only,
-component generation, not micro-tweaks.
+not micro-tweaks.
 
 ## graphify
 
-This project has a knowledge graph at graphify-out/ with god nodes, community structure, and cross-file relationships.
+Knowledge-graph navigation via graphify CLI. ALL rules conditional:
+`graphify-out/graph.json` exists in the project — else skip graphify
+entirely, read files directly.
 
-Rules:
+- Codebase-wide question → `graphify query "<question>"` first.
-- For codebase questions, first run `graphify query "<question>"` when graphify-out/graph.json exists. Use `graphify path "<A>" "<B>"` for relationships and `graphify explain "<concept>"` for focused concepts. These return a scoped subgraph, usually much smaller than GRAPH_REPORT.md or raw grep output.
+  Relationships → `graphify path "<A>" "<B>"`. Focused concept →
-- If graphify-out/wiki/index.md exists, use it for broad navigation instead of raw source browsing.
+  `graphify explain "<concept>"`. Scoped subgraph beats raw grep.
-- Read graphify-out/GRAPH_REPORT.md only for broad architecture review or when query/path/explain do not surface enough context.
+- Known file/symbol lookup, small task (hotfix, typo, single file) →
-- After modifying code, run `graphify update .` to keep the graph current (AST-only, no API cost).
+  read directly, no graphify.
+- `graphify-out/wiki/index.md` exists → broad-navigation entrypoint.
+  `GRAPH_REPORT.md` only for whole-architecture review or when
+  query/path/explain insufficient.
+- After modifying code → `graphify update .` (AST-only, no API cost).
+
+---
+
+# This repo only (claude-config)
+
+Apply when working directory = the claude-config repo itself.
+
+## Health Stack
+- shell: `shellcheck *.sh hooks/*.sh lib/*.sh`