Run all quality checks, review changes against the current issue, perform code review, and create a PR.
Remove signs of AI-generated writing from text. Use when editing or reviewing text to make it sound more natural and human-written. Based on Wikipedia's comprehensive "Signs of AI writing" guide. Detects and fixes patterns including: inflated symbolism, promotional language, superficial -ing analyses, vague attributions, em dash overuse, rule of three, AI vocabulary words, passive voice, negative parallelisms, and filler phrases.
Execute (or resume) an implementation plan stored in .implementation-plan.md. Use when the user types /implement-plan, asks to "resume implementation", or "continue the plan". Reads .implementation-plan.md, picks up at the first uncompleted step in the Execution Progress checklist, and dispatches a Sonnet agent to finish the work.
End-to-end GitHub issue implementation — /implement <N>. Fetches the issue via GitHub CLI, writes it locally, then either implements directly (easy issues) or plans with Opus and implements with Sonnet (complex issues). Resilient to interruption — resumes from the last checkpoint. Also triggered by "implement issue
Create or refine GitHub issues — /issue <description>, /issue --r <N> [guidance] (Refine Mode), /issue --f <description> (Fast Mode, skips clarifying questions and persona review). Parses intent, validates against project goals, invokes persona agents, and handles decomposition. Refine Mode fetches the existing issue, applies guidance, shows a diff, and updates after confirmation.
Lint and auto-fix agent-only markdown files for token efficiency and model readability.