Files
01e15490f0 fix(skill-evolution): wire Skill PostToolUse tracker so skill-health shows real runs (#2490)
* fix(skill-evolution): wire Skill PostToolUse tracker so skill-health shows real runs (#2463)

recordSkillExecution() had no production callers, so
~/.claude/state/skill-runs.jsonl was never written and
`scripts/skills-health.js --dashboard` always reported 0 runs.

Adds scripts/hooks/skill-run-tracker.js and registers it as an async
PostToolUse hook (matcher: Skill) in posttooluse-dispatcher.js, which is now
the single PostToolUse entrypoint on main.

Addresses the privacy and bounds review on #2555's sibling PR:

- No prompt text is persisted. task_description is synthesized as
  "Skill invocation: <skill_id>"; tool_input.task_description/description/
  prompt are never read.
- Every persisted string is bounded and charset-restricted. A skill id is an
  identifier, so free text, newlines, or an over-long value are dropped rather
  than truncated and written through.
- The JSONL sink is created 0600 and re-tightened on each append, repairing
  files written before this bound existed.
- The sink is capped at MAX_RUN_RECORDS (5000), trimmed oldest-first, so the
  append-only file can no longer grow without limit.

Tests cover the privacy guarantee (no prompt text reaches a record), the
identifier bounds, the file mode on POSIX, and the retention cap.

* fix(skill-evolution): re-register the tracker for PostToolUseFailure

The rebase onto current main dropped the hooks.json entry, which silently
resurrected the P1 from the earlier review round: deriveOutcome() still
branches on hook_event_name === 'PostToolUseFailure', but the PostToolUse
dispatcher does not fan that event out, so the branch was unreachable in
production. Hard Skill failures were dropped from telemetry entirely, which
inflates the dashboard success rate — the opposite of what #2463 asks for.

Restores the dedicated PostToolUseFailure entry (matcher Skill, id
post:skill:track, same run-with-flags wrapper and standard,strict gating as
the dispatcher registration). Verified end-to-end: a PostToolUseFailure
payload piped through run-with-flags now records outcome "failure".

Adds a regression test asserting the registration so a future rebase cannot
quietly drop it again.

---------

Co-authored-by: haelyra <49814733+haelyra@users.noreply.github.com>
2026-08-12 18:12:49 -04:00
..

Hooks

Hooks are event-driven automations that fire before or after Claude Code tool executions. They enforce code quality, catch mistakes early, and automate repetitive checks.

How Hooks Work

User request → Claude picks a tool → PreToolUse hook runs → Tool executes → PostToolUse hook runs
  • PreToolUse hooks run before the tool executes. They can block (exit code 2) or warn (stderr without blocking).
  • PostToolUse hooks run after the tool completes. They can analyze output but cannot block.
  • Stop hooks run after each Claude response.
  • SessionStart/SessionEnd hooks run at session lifecycle boundaries.
  • PreCompact hooks run before context compaction, useful for saving state.

Hooks in This Plugin

Memory persistence lifecycle definitions live in hooks/memory-persistence/. The executable hook graph remains hooks/hooks.json; the memory persistence directory is the stable contract for SessionStart, PreCompact, observation, activity tracking, and SessionEnd behavior.

Installing These Hooks Manually

For Claude Code manual installs, do not paste the raw repo hooks.json into ~/.claude/settings.json or copy it directly into ~/.claude/hooks/hooks.json. The checked-in file is plugin/repo-oriented and is meant to be installed through the ECC installer or loaded as a plugin.

Use the installer instead so hook commands are rewritten against your actual Claude root:

bash ./install.sh --target claude --modules hooks-runtime
pwsh -File .\install.ps1 --target claude --modules hooks-runtime

That installs resolved hooks to ~/.claude/hooks/hooks.json. On Windows, the Claude config root is %USERPROFILE%\\.claude.

PreToolUse Hooks

Hook Matcher Behavior Exit Code
Dev server blocker Bash Blocks npm run dev etc. outside tmux — ensures log access 2 (blocks)
Tmux reminder Bash Suggests tmux for long-running commands (npm test, cargo build, docker) 0 (warns)
Git push reminder Bash Reminds to review changes before git push 0 (warns)
Pre-commit quality check Bash Runs quality checks before git commit: lints staged files, validates commit message format when provided via -m/--message, detects console.log/debugger/secrets 2 (blocks critical) / 0 (warns)
Doc file warning Write Warns about non-standard .md/.txt files (allows README, CLAUDE, CONTRIBUTING, CHANGELOG, LICENSE, SKILL, docs/, skills/); cross-platform path handling 0 (warns)
Strategic compact Edit|Write Suggests manual /compact at logical intervals (every ~50 tool calls) 0 (warns)

PostToolUse Hooks

Hook Matcher What It Does
PR logger Bash Logs PR URL and review command after gh pr create
Build analysis Bash Background analysis after build commands (async, non-blocking)
Quality gate Edit|Write|MultiEdit Runs fast quality checks after edits
Design quality check Edit|Write|MultiEdit Warns when frontend edits drift toward generic template-looking UI
Prettier format Edit Auto-formats JS/TS files with Prettier after edits
TypeScript check Edit Runs tsc --noEmit after editing .ts/.tsx files
console.log warning Edit Warns about console.log statements in edited files

Lifecycle Hooks

Hook Event What It Does
Session start SessionStart Loads previous context and detects package manager
Plan Canvas sessions SessionStart Surfaces open Plan Canvas browser reviews so a fresh session can resume the loop
Pre-compact PreCompact Saves state before context compaction
Console.log audit Stop Checks all modified files for console.log after each response
Session summary Stop Persists session state when transcript path is available
Pattern extraction Stop Evaluates session for extractable patterns (continuous learning)
Cost tracker Stop Emits lightweight run-cost telemetry markers
Desktop notify Stop Sends macOS desktop notification with task summary (standard+)
Session end marker SessionEnd Lifecycle marker and cleanup log

Customizing Hooks

Disabling a Hook

Remove or comment out the hook entry in hooks.json. If installed as a plugin, override in your ~/.claude/settings.json:

{
  "hooks": {
    "PreToolUse": [
      {
        "matcher": "Write",
        "hooks": [],
        "description": "Override: allow all .md file creation"
      }
    ]
  }
}

Use environment variables to control hook behavior without editing hooks.json:

# Master switch. Explicit environment values override plugin preferences.
export ECC_HOOKS_ENABLED=true

# minimal | standard | strict (default: standard)
export ECC_HOOK_PROFILE=standard

# Disable specific hook IDs (comma-separated)
export ECC_DISABLED_HOOKS="pre:bash:tmux-reminder,post:edit:typecheck"

# Disable only GateGuard during setup or recovery
export ECC_GATEGUARD=off

# Cap SessionStart additional context (default: 8000 chars)
export ECC_SESSION_START_MAX_CHARS=4000

# Disable SessionStart additional context entirely
export ECC_SESSION_START_CONTEXT=off

# Keep context/scope/loop warnings but suppress API-rate cost estimates
export ECC_CONTEXT_MONITOR_COST_WARNINGS=off

Windows PowerShell:

[Environment]::SetEnvironmentVariable('ECC_CONTEXT_MONITOR_COST_WARNINGS', 'off', 'User')

Claude setup-only value:

  • off — disables local ECC hook work through ecc setup; it is not a runtime hook profile.

Runtime hook profiles:

  • minimal — keep essential lifecycle and safety hooks only.
  • standard — default; balanced quality + safety checks.
  • strict — enables additional reminders and stricter guardrails.

The Claude plugin exposes the same choices as the personal hooks_enabled and hook_profile settings. Run ecc setup --mode claude-plugin to install or update the plugin and change those preferences.

Writing Your Own Hook

Hooks are shell commands that receive tool input as JSON on stdin and must output JSON on stdout.

Basic structure:

// my-hook.js
let data = '';
process.stdin.on('data', chunk => data += chunk);
process.stdin.on('end', () => {
  const input = JSON.parse(data);

  // Access tool info
  const toolName = input.tool_name;        // "Edit", "Bash", "Write", etc.
  const toolInput = input.tool_input;      // Tool-specific parameters
  const toolOutput = input.tool_output;    // Only available in PostToolUse

  // Warn (non-blocking): write to stderr
  console.error('[Hook] Warning message shown to Claude');

  // Block (PreToolUse only): exit with code 2
  // process.exit(2);

  // Always output the original data to stdout
  console.log(data);
});

Exit codes:

  • 0 — Success (continue execution)
  • 2 — Block the tool call (PreToolUse only)
  • Other non-zero — Error (logged but does not block)

Hook Input Schema

interface HookInput {
  tool_name: string;          // "Bash", "Edit", "Write", "Read", etc.
  tool_input: {
    command?: string;         // Bash: the command being run
    file_path?: string;       // Edit/Write/Read: target file
    old_string?: string;      // Edit: text being replaced
    new_string?: string;      // Edit: replacement text
    content?: string;         // Write: file content
  };
  tool_output?: {             // PostToolUse only
    output?: string;          // Command/tool output
  };
}

Async Hooks

For hooks that should not block the main flow (e.g., background analysis):

{
  "type": "command",
  "command": "node my-slow-hook.js",
  "async": true,
  "timeout": 30
}

Async hooks run in the background. They cannot block tool execution.

Common Hook Recipes

Warn about TODO comments

{
  "matcher": "Edit",
  "hooks": [{
    "type": "command",
    "command": "node -e \"let d='';process.stdin.on('data',c=>d+=c);process.stdin.on('end',()=>{const i=JSON.parse(d);const ns=i.tool_input?.new_string||'';if(/TODO|FIXME|HACK/.test(ns)){console.error('[Hook] New TODO/FIXME added - consider creating an issue')}console.log(d)})\""
  }],
  "description": "Warn when adding TODO/FIXME comments"
}

Block large file creation

{
  "matcher": "Write",
  "hooks": [{
    "type": "command",
    "command": "node -e \"let d='';process.stdin.on('data',c=>d+=c);process.stdin.on('end',()=>{const i=JSON.parse(d);const c=i.tool_input?.content||'';const lines=c.split('\\n').length;if(lines>800){console.error('[Hook] BLOCKED: File exceeds 800 lines ('+lines+' lines)');console.error('[Hook] Split into smaller, focused modules');process.exit(2)}console.log(d)})\""
  }],
  "description": "Block creation of files larger than 800 lines"
}

Auto-format Python files with ruff

{
  "matcher": "Edit",
  "hooks": [{
    "type": "command",
    "command": "node -e \"let d='';process.stdin.on('data',c=>d+=c);process.stdin.on('end',()=>{const i=JSON.parse(d);const p=i.tool_input?.file_path||'';if(/\\.py$/.test(p)){const{execFileSync}=require('child_process');try{execFileSync('ruff',['format',p],{stdio:'pipe'})}catch(e){}}console.log(d)})\""
  }],
  "description": "Auto-format Python files with ruff after edits"
}

Require test files alongside new source files

{
  "matcher": "Write",
  "hooks": [{
    "type": "command",
    "command": "node -e \"const fs=require('fs');let d='';process.stdin.on('data',c=>d+=c);process.stdin.on('end',()=>{const i=JSON.parse(d);const p=i.tool_input?.file_path||'';if(/src\\/.*\\.(ts|js)$/.test(p)&&!/\\.test\\.|\\.spec\\./.test(p)){const testPath=p.replace(/\\.(ts|js)$/,'.test.$1');if(!fs.existsSync(testPath)){console.error('[Hook] No test file found for: '+p);console.error('[Hook] Expected: '+testPath);console.error('[Hook] Consider writing tests first (/tdd)')}}console.log(d)})\""
  }],
  "description": "Remind to create tests when adding new source files"
}

Cross-Platform Notes

Hook logic is implemented in Node.js scripts for cross-platform behavior on Windows, macOS, and Linux. The continuous-learning observer is exposed as a Node-mode hook and delegates to its existing observe.sh implementation through a profile-gated runner with Windows-safe fallback behavior.