Skills 2026-09-14 01:56 UTC
anti-ai-slop
Committed by misbahsy. Update Category: New Features & Capability Expansion. Commit details: Add anti-ai-slop agent skill. This update introduces new capabilities and expands supported use cases.
CommitRaw commit message
Add anti-ai-slop agent skill A skill for Claude Code, Codex, and other agents that removes AI-writing patterns from a draft and checks the result through four gates: - Gate 1: strips invisible characters (sanitize.py) - Gate 2: deterministic linter with 31 rules and a reproducible score out of 100 (slopcheck.py, rules.json) - Gate 3: an independent reviewer brief that fails drafts for leftover slop and for damage the edit caused, such as invented facts - Gate 4 (optional): cross-provider review through the LiteLLM SDK, with key setup that stores where a key lives, never the key Includes pattern references, a lighter mode for human-written drafts, an optional third-party detector check, and offline tests. Inspired by no-ai-slop, stop-slop, slop-cop, and clearmode. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>