chore: customize skill graph for socket-cli

Author: jdalton

.claude/agents/code-reviewer.md

@@ -1,4 +1,4 @@
-You are a code reviewer for a Node.js/TypeScript monorepo (socket-registry).
+You are a code reviewer for a Node.js/TypeScript monorepo (socket-cli).
 
 Apply the rules from CLAUDE.md sections listed below. Reference the full section in CLAUDE.md for details — these are summaries, not the complete rules.
 

.claude/agents/refactor-cleaner.md

@@ -1,4 +1,4 @@
-You are a refactoring specialist for a Node.js/TypeScript monorepo (socket-registry).
+You are a refactoring specialist for a Node.js/TypeScript monorepo (socket-cli).
 
 Apply these rules from CLAUDE.md exactly:
 

.claude/commands/quality-loop.md

@@ -1,55 +1,21 @@
-# quality-loop
+Run the `/quality-scan` skill and fix all issues found. Repeat until zero issues remain or 5 iterations complete.
 
-Run comprehensive quality scan and automatic issue fixing loop for socket-cli.
-
-## What it does
-
-Executes an iterative quality improvement cycle:
-
-1. Updates dependencies
-2. Cleans up repository (removes junk files)
-3. Validates code structure
-4. Runs specialized quality scans
-5. Fixes ALL issues found
-6. Commits fixes
-7. Repeats until zero issues or 5 iterations
-
-## Usage
-
-```bash
-/quality-loop
-```
-
-## Scan types
-
-- **critical** - Crashes, security, data corruption, auth handling
-- **logic** - Algorithm errors, edge cases, validation bugs
-- **cache** - Config/token caching correctness
-- **workflow** - Build scripts, CI/CD, cross-platform compatibility
-- **security** - GitHub Actions security, credential handling
-- **documentation** - Command examples, flag accuracy, API docs
+**Interactive only** — this command makes code changes and commits. Do not use as an automated pipeline gate.
 
 ## Process
 
-The skill will:
-- Ask which scans to run (default: all)
-- Run dependency updates
-- Clean junk files with confirmation
-- Execute selected scans sequentially
-- Aggregate and deduplicate findings
-- Fix issues and commit changes
-- Repeat until clean or max iterations
-
-## Exit conditions
-
-- ✅ Success: Zero issues found
-- ⏹️ Stop: After 5 iterations (prevent infinite loops)
-
-## Notes
-
-- Commits fixes with proper git messages
-- Skips no issues (fixes architectural problems too)
-- Runs tests after each iteration
-- Reports progress and statistics
-
-Use this command to maintain high code quality standards across socket-cli.
+1. Run `/quality-scan` skill (all scan types)
+2. If issues found: spawn the `refactor-cleaner` agent (see `agents/refactor-cleaner.md`) to fix them, grouped by category
+3. Run verify-build (see `_shared/verify-build.md`) after fixes
+4. Run `/quality-scan` again
+5. Repeat until:
+   - Zero issues found (success), OR
+   - 5 iterations completed (stop)
+6. Commit all fixes: `fix: resolve quality scan issues (iteration N)`
+
+## Rules
+
+- Fix every issue, not just easy ones
+- Spawn refactor-cleaner with CLAUDE.md's pre-action protocol: dead code first, then structural changes, ≤5 files per phase
+- Run tests after fixes to verify nothing broke
+- Track iteration count and report progress

.gitignore

@@ -74,11 +74,11 @@ yarn-error.log*
 # ============================================================================
 *.log
 **/build/*.log
-**/.claude/*
-!**/.claude/agents/
-!**/.claude/commands/
-!**/.claude/ops/
-!**/.claude/skills/
+/.claude/*
+!/.claude/agents/
+!/.claude/commands/
+!/.claude/ops/
+!/.claude/skills/
 
 # ============================================================================
 # Backup and temporary files

CLAUDE.md

@@ -77,7 +77,7 @@
 - Always prefer editing existing files
 - Forbidden to create docs unless requested
 - Required to do exactly what was asked
-- NEVER use `npx`, `pnpm dlx`, or `yarn dlx` — use `pnpm exec` or `pnpm run` with pinned devDeps
+- 🚨 **NEVER use `npx`, `pnpm dlx`, or `yarn dlx`** — use `pnpm exec <package>` for devDep binaries, or `pnpm run <script>` for package.json scripts. If a tool is needed, add it as a pinned devDependency first.
 
 ## ROLE
 
@@ -497,14 +497,12 @@ The file system is working memory. Use it actively:
 
 ## Agents & Skills
 
-Reusable automation infrastructure lives in `.claude/`:
-
-- **Agents** (`.claude/agents/`): Specialized sub-agent definitions (code-reviewer, security-reviewer, refactor-cleaner)
-- **Skills** (`.claude/skills/`): Multi-phase pipelines with shared components in `_shared/`
-- **Ops** (`.claude/ops/queue.yaml`): Pipeline run tracking and phase ordering
-- **Commands** (`.claude/commands/`): Slash-command entry points that invoke skills
-
-Skills reference shared components (`env-check`, `verify-build`, `security-tools`, `report-format`) and can spawn agents for specialized work. Pipeline state is tracked in `queue.yaml`.
+- `/security-scan` — runs AgentShield + zizmor security audit
+- `/quality-scan` — comprehensive code quality analysis
+- `/quality-loop` — scan and fix iteratively
+- Agents: `code-reviewer`, `security-reviewer`, `refactor-cleaner` (in `.claude/agents/`)
+- Shared subskills in `.claude/skills/_shared/`
+- Pipeline state tracked in `.claude/ops/queue.yaml`
 
 ## Quality Standards