HTTPX Humanize

•
16
•

Eliminates AI writing tells, filler phrasing, and robotic scaffolding from documentation, pull request comments, and prose without altering code.

HTTPX Humanize refines AI-generated documentation, pull request comments, and conversational prose so they read naturally like an experienced engineer wrote them. It systematically strips obvious LLM artifacts—such as em-dashes, repetitive filler openers, corporate buzzwords, excessive apologies, and chatbot conversational scaffolding—while preserving technical substance and leaving code blocks, identifiers, and syntax completely untouched. By combining context-aware rewriting rules with Klaussy's deterministic backstop scrubber, the skill ensures concise, direct, and authentic communication across markdown documentation, commit messages, and team reviews.

主な機能

01Filters out LLM buzzwords, corporate jargon, and overused transition crutches
02Formats PR comments and review replies into concise, one- to five-sentence engineer prose
03Integrates with Klaussy's deterministic scrubber CLI as a reliable post-processing backstop
04Preserves all inline code, fenced code blocks, and technical identifiers without changes
0516 GitHub stars
06Strips distinct AI tells including em-dashes, chatbot pleasantries, and throat-clearing openers

ユースケース

01Cleaning up AI-generated project documentation, guides, and markdown files to sound natural
02Rewriting verbose codebase comments into clear, concise, load-bearing explanations
03Polishing PR review comments and automated review replies before posting to GitHub or GitLab