Local-first, evidence-controlled academic writing workflows for AI agents, with bounded revision, clean-room review, and release governance.

Open-source agent skills for AI for Science: topic exploration, literature survey, experiments, paper writing, and integrity audit — driven by any coding agent.