Monitor and steer Claude Code, Codex, or other terminal coding agents running in tmux panes. Use when asked to supervise, babysit, drive, or automate agent sessions through tmux, including reading agent output, detecting idle/blocked/context-boundary states, sending approved commands, clearing or compacting context at explicit boundaries, and coordinating multi-agent workflows without duplicating Claude/Codex-specific skill content.
Multi-provider interactive PR code review. Runs Claude Code, Gemini CLI, and Codex CLI in parallel to review a pull request, merges and deduplicates findings by severity, then walks through each issue interactively so you can vet, discuss, and selectively post comments. Use this skill when the user wants to review a PR, asks for code review on a pull request, says 'review PR #123', 'code review this PR', 'check this PR for issues', or mentions reviewing changes before merging. Also triggers on 'multi-provider review', 'cross-AI review', or 'security review PR'.
Analyze the current session transcript and improve project documentation based on what was learned. Triggered automatically via SessionEnd hook, or manually with /reflect. Use when the user says "reflect", "what did we learn", "improve docs from this session", or "capture learnings". Also runs automatically in the background after every substantial session.