Search and filter by type across the catalog
170 results in Hooks, Memory · page 3 of 8
Automatically loads AGENTS.md configuration file content at session start to ensure Claude Code follows project-specific agent behavior. Only loads if AGENTS.md exists, otherwise passes empty context. Supports the universal AGENTS.md standard for cross-platform AI assistant compatibility.
AI-powered bash command security guard. Before any Bash command runs, a lightweight Claude subagent evaluates it for destructive or irreversible patterns — recursive deletes, force pushes to protected branches, database drops, and credential exposure — and blocks execution with a clear explanation if flagged. Uses PreToolUse with type:agent, which is the only hook pattern that can block tool execution via AI reasoning.
Create automatic backup of files before any Edit operation for safety. This hook creates a timestamped backup copy (filename.backup.timestamp) of any existing file before Claude modifies it. Provides a safety net to recover previous versions if needed. Only backs up existing files, includes error suppression to handle edge cases gracefully.
Real-time browser visualization of Claude Code's context window and subagent/tool execution as a git-graph timeline. Opens http://localhost:7878 on SessionStart with a vertical timeline (most recent on top), one column per agent (main + subagents branching from their Task tool call). Tool calls are color-coded by type (Read=green, Edit/Write=orange, Bash=red, Grep/Glob=cyan, Task=purple, Web=yellow, MCP=gray). Right sidebar shows context-window usage (tokens/200K) with cache_read/cache_creation/input/output breakdown plus per-subagent mini-context. Reads ~/.claude/projects/<encoded-cwd>/<session_id>.jsonl directly — no data replication, no network calls. Pure stdlib Python, zero pip dependencies. Persistent daemon HTTP server on port 7878 (auto-fallback 7879-7888 if busy; override with CONTEXT_TIMELINE_PORT env var). Watchdog auto-shutdown after 1h of inactivity. Disable browser auto-open with CONTEXT_TIMELINE_NO_BROWSER=1. Manual shutdown: python3 .claude/hooks/context-timeline.py --shutdown
context-timeline: Claude Code session visualizer. Modes (called by Claude Code hooks via stdin JSON): --server-start SessionStart hook — launch daemon + open browser --event <NAME> PreToolUse / PostToolUse / Stop hook — notify server --shutdown Kill daemon manually --run-server <PORT> Internal: daemon entry point (do not call directly)
Enforce conventional commit message format for all git commits. Validates commit messages follow the pattern: type(scope): description. Supported types: feat, fix, docs, style, refactor, perf, test, chore, ci, build, revert. Ensures consistent commit history for changelog generation and semantic versioning.
❌ Invalid commit message format Your message: {commit_msg} Commit messages must follow Conventional Commits: type(scope): description Types: feat: New feature fix: Bug fix docs: Documentation changes style: Code style changes (formatting) refactor: Code refactoring perf: Performance improvements test: Adding or updating tests chore: Maintenance tasks ci: CI/CD changes build: Build system changes revert: Revert previous commit Examples: ✅ feat: add user authentication ✅ feat(auth): implement JWT tokens ✅ fix: resolve memory leak in parser ✅ fix(api): handle null responses ✅ docs: update API documentation Invalid: ❌ Added new feature (no type) ❌ feat:add feature (missing space after colon) ❌ feature: add login (wrong type, use 'feat') 💡 Tip: Start your message with one of the types above followed by a colon and space.
Advanced protection against dangerous shell commands with multi-level security. Blocks catastrophic operations (rm -rf /, dd, mkfs), protects critical paths (.claude/, .git/, node_modules/), and warns about suspicious patterns. Features: catastrophic command blocking, critical path protection, smart pattern detection, and detailed safety messages.
Auto Debug Log Viewer. Opens a live-tailing debug log window when Claude Code starts with --debug or -d flag. The window closes automatically on session end. To keep the debug window open after session ends, set DEBUG_WINDOW_AUTO_CLOSE_DISABLE=1 in your settings.json. Tested on Intel Mac. Supports macOS, Linux, and Windows (Git Bash/Cygwin). Contributions from other platform users are welcome.
Monitor deployment status, error rates, and performance metrics, sending notifications for failed deployments or performance degradation. Tracks Vercel deployment health, monitors build success/failure rates, and provides alerts for deployment issues. Setup: Export 'export VERCEL_TOKEN=your_token' and 'export VERCEL_PROJECT_ID=your_project_id' (get from vercel.com/account/tokens and Vercel dashboard).
Sends a native desktop notification when Claude Code finishes responding. Uses the Stop hook event so you get a single notification per response instead of one per tool call (which is very noisy with PostToolUse). Supports macOS (osascript) and Linux (notify-send). Useful when you switch to another window while Claude works — you'll get a notification when it's ready for your input.
Send detailed Discord notifications with session information when Claude Code finishes. Includes working directory, session duration, and system info with rich embeds. Requires DISCORD_WEBHOOK_URL environment variable.