ai-agent-rules

Performs thorough code review on local changes or PRs. Use this skill proactively after implementing code changes to catch issues before commit/push. Also use when reviewing PRs from other engineers.

Generate a portable session handoff document (task, decisions, state, next steps) as a single copy-pasteable markdown block for resuming on another device or terminal.

Create or update AGENTS.md with repo-specific patterns, conventions, commands, and gotchas for LLM coding agents. Use after implementing features that introduce new patterns worth documenting.

Continues after a session crash while working on a task

Get multi-model perspectives (Codex + Gemini) on any work product. Use at any development stage — planning, design, implementation, testing — to catch blind spots the primary agent might miss.

Creates or updates PLAN.md based on session - auto-detects create vs update mode

Write, update, or review documentation (README, ARCHITECTURE.md, API docs, guides). Use after implementing features to document new APIs, CLI commands, or behavior changes.

Creates GitHub pull requests with comprehensive descriptions by analyzing git history and code changes

Analyze your historical Claude Code prompts and get personalized feedback on how to improve your prompting technique

Expert guidance for writing and optimizing LLM prompts. Use when creating or updating AGENTS.md, CLAUDE.md, SKILL.md, system prompts, or custom instructions.

Rebase a branch onto main (or a target) with squash-first strategy, semantic merge verification, and safe force-push.

Multi-agent research on any topic. Use when the user asks to research, investigate, or deep-dive a question requiring multiple sources. Scales from inline answers to 10 parallel research agents based on complexity.

Search previous agent sessions and conversations (Claude Code, Codex, Gemini, Goose, Amp, Buzz). Use to find session transcripts, recover prior outputs, or compare recent sessions for a repository.

Write, update, or review tests. Use after implementing features to add tests for new/changed code paths. Covers testing strategy, framework selection, and coverage analysis.

Clean up stale git worktrees to reclaim disk space. Detects merged branches, deleted remotes, and abandoned worktrees. Rust-aware (reports target/ sizes separately). Use for worktree cleanup or disk reclamation.