papr

Designs, verifies, runs, and monitors experiments from reviewer feedback, with TDD dry-runs and figure generation. Use when the user says "design experiments", "run ablation", "need new results", or when a papr-panel action list contains "requires new experiment" items.

Runs a 6-agent parallel paper review panel via Agent Teams (live cross-talk). Agents: Advisor (storyline), Expert (technical+baselines), Standard (quality), Brief (main-text only), Lay (accessibility), Author (defense+action list). Use when the user says "review my paper", "run panel", "get feedback", or asks for multi-perspective critique. Requires CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS=1.

Implements text changes to a LaTeX paper, downloads real bibtex for new citations, removes AI writing patterns via /humanizer, and compiles the PDF. Use when the user says "apply changes", "fix the paper", "implement feedback", or after a review panel produces an action list.

Runs an autonomous multi-round paper improvement loop on a LaTeX project, chaining /papr-panel -> /papr-experiment -> /papr-write -> /papr-review per round and stopping at score 9+/10. Use when the user says "improve my paper", "review pipeline", "run papr", or asks for iterative paper improvement. Heavy side effects: long compute, latex compiles, external Codex calls.

Sends a paper for external blind review via the Codex plugin (latest GPT model) and compares external findings with the internal panel. Use when the user says "external review", "get another opinion", "blind review", or after papr-write completes.