Use when the user wants the current code proactively reviewed and fixed (e.g. "review-loop", "/review-loop", "review what I have and fix it"). Self + cross-model review → triage → fix → update PR, loop until clean.
Use when the user wants work packaged for review (e.g. "review packet", "/review-packet", "prepare this for review", "make it easy to review this PR set / design / research"). Builds a decision-first page that pairs every claim with the evidence for it.
Use when the user asks for a cross-review by Claude (e.g. types "claude-review", "/claude-review", or "have Claude review this"). Invokes Claude Code headlessly to review the current branch/changes, prints Claude's findings verbatim, then triages them…
Use when the user asks for a diagram (architecture, flow, sequence) as an image. Writes an Excalidraw file and exports it to PNG.
Use when the user says gpra or asks to rebase the current branch over the remote default branch with autostash. Never resolves conflicts automatically and only force-pushes after confirmation.
Use when the user asks for a session retrospective or retro, or how to avoid friction from this session in future. Reviews the session and proposes concrete AGENTS.md, config, skill, and doc changes, applying only the ones the user approves.
Use when the user asks to take a task end-to-end autonomously (e.g. "autodev", "/autodev", "build this on autopilot"). Runs the full chain — plan → implement → review → PR → converge — stopping only at real decisions.
Use when the user asks for a handoff, a prompt for another agent, or to delegate a task to a fresh session (Claude, Codex, a teammate). Writes a self-contained discussion prompt or work order and copies it to the clipboard.