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

Open-source Agent Skill that scores AI-sounding writing 0–100, edits it with your assistant.