What problem does it solve? AI-generated text leaks into developer artifacts—docs, docstrings, commit messages, and PR descriptions—bringing promotional language, chat leaks, tautological docstrings, and filler that erode trust and clarity. This Skill scans those artifacts and reports each pattern with a concrete fix suggestion. ## Core Features & Use Cases - Six pattern categories: Detects content, vocabulary, formatting, communication, filler, and code-docs issues using a detailed reference catalog with false-positive rules. - Scoped scanning: Reviews changed files from main by default, or the full codebase with --all, partitioned into Prose, Code Docs, and Git artifact groups. - Structured reporting: Writes a validated ai-writing-review.json report with risk and fix-safety ratings, plus a human-readable summary grouped by category. - Use Case: Before merging a PR, run the review to catch commit messages starting with "Certainly!" or docstrings that merely restate function names, then hand findings to a humanize pass. ## Quick Start Ask the agent to review the current branch's changed files for AI-generated writing patterns and produce a findings report.
