core

Background agent that monitors CI/CD pipelines for open PRs/MRs and automatically triages and fixes failures. Spawns isolated fixer agents in worktrees to resolve lint, format, and test failures without disrupting the main working tree.

Design system architecture and high-level technical strategy

Use when multiple fix attempts fail and you need to systematically restore to a working baseline and reimplement instead of fixing broken code.

Use when modifying existing files, refactoring, improving code quality, or touching legacy code by applying the Boy Scout Rule to leave code better than you found it.

Code review a pull request

Investigate and diagnose issues without necessarily fixing them

Guides end-to-end feature development through 8 phases: discover requirements, explore codebase patterns, clarify ambiguities with the user, design architecture, implement with TDD, run multi-agent code review, validate all quality gates, and write a blog post. Use when asked to add a feature, implement a new capability, build functionality, or develop a feature end-to-end.

Generate or update documentation for code, APIs, or systems

Explain code, concepts, or technical decisions in clear, understandable terms

Prefer explicit configuration over framework defaults to prevent environment-dependent failures

Debug and fix bugs, errors, or unexpected behavior

Use when modifying, removing, or refactoring code that lacks test coverage. Emphasizes the danger of untested changes and the RGR workflow to add characterization tests before modifications.

Optimize code for performance, readability, or efficiency

Use when designing modules, APIs, and system architecture requiring independent, non-overlapping components where changes in one don't affect others.

Create tactical implementation plan for a feature or task

Use when responding to questions or providing information requiring professional honesty and directness over excessive agreeableness.

Use when setting up or organizing Claude Code project memory (CLAUDE.md, .claude/rules/) for better context awareness, consistent behavior, and project-specific instructions.

Use automatically during development workflows when making claims about tests, builds, verification, or code quality requiring concrete evidence to ensure trust through transparency.

Restructure code to improve quality without changing behavior

Multi-agent code review with confidence-based filtering

Reduce token waste by batching tool calls and avoiding narration between sequential operations

Use when designing solutions, adding features, or refactoring by applying KISS, YAGNI, and Principle of Least Astonishment to write simple, predictable code.

Use during implementation when designing modules, functions, and components requiring SOLID principles for maintainable, flexible architecture.

Use when designing modules and components requiring Composition Over Inheritance, Law of Demeter, Tell Don't Ask, and Encapsulation principles that transcend programming paradigms.

Write tests for code using test-driven development principles

Minimize token consumption through efficient tool usage patterns

18 hooks across 11 events