SlopMonster

SlopMonster checks English drafts for wording and sentence patterns often associated with AI writing. It's for writers and developers reviewing copy for READMEs, landing pages, emails or scripts.

• Pattern checks: The checker flags listed vocabulary, formulaic constructions, dense punctuation and certain three-item lists, then reports matches with a score out of five.
• Claim warnings: It marks numbers near words such as "users" or "customers" for review, but can't verify those claims or their evidence.
• Agent editing guidance: Claude Code, Codex and other agents can follow its instructions to review a draft's purpose, audience, structure and factual support.
• Optional second review: A separate editing pass uses a different model family and requires an installed, authenticated AI command-line tool, using that tool's configured model.
• Repository checks: Teams can include the checker in automated checks and fail a workflow when it finds scored patterns; blocking a merge depends on the repository's rules.

The standalone checker runs locally with Python 3 and its standard library. It needs no API key or extra packages. The project is open source under the MIT license. The optional model-based review depends on the chosen AI tool and isn't part of the standalone check.

A passing score means no scored patterns matched, not that the writing is good. The checker can't identify an author or judge factual accuracy. Its rules target English, so text in another language can pass simply because the patterns don't match.