AI Writing Reviewer
Detects and flags AI-generated writing patterns across documentation, code comments, and git messages to ensure authentic developer communication.
AI Writing Reviewer is a specialized Claude Code skill designed to maintain the authenticity and clarity of technical artifacts. It scans Markdown files, source code docstrings, and Git commit histories to identify characteristic LLM patterns such as 'chat leaks,' repetitive vocabulary, and tautological documentation. By parallelizing scans across artifact groups, it provides actionable suggestions to replace robotic or promotional language with human-centric, informative text, helping teams maintain high standards for their project's written communication.
主な機能
0163 GitHub stars
02Structured JSON reporting with specific improvement suggestions and risk assessments
03Identification of tautological docstrings that repeat function names without adding value
04Support for parallelized scanning via subagents for efficient processing of large codebases
05Automated detection of AI-generated vocabulary and 'chat leaks' in PRs and commits
06Cross-artifact scanning covering Markdown, source code, and Git history
ユースケース
01Auditing project documentation to remove robotic or promotional filler text
02Ensuring code docstrings provide meaningful technical context rather than just restating function names
03Reviewing Git commit messages and PR descriptions for accidental LLM 'chat' artifacts