Install
openclaw skills install @mwclaw/anti-ai-slop-writing-gateDetect or minimally edit formulaic AI-style prose while preserving the writer's voice.
openclaw skills install @mwclaw/anti-ai-slop-writing-gateUse this skill to edit a draft, identify checkable AI-slop patterns without rewriting it, or run a silent pre-send gate on substantive prose.
Do not use stylistic patterns to claim that AI wrote a text. The skill evaluates the writing, not its authorship.
Use pre-send mode for interpretations, advice, research synthesis, consequential explanations, and other developed replies. Skip it for acknowledgments, exact factual one-liners, raw tool output, or urgent safety instructions.
Clean VersionWhat Changed — short and specificRemaining Risk only when meaning, sourcing, or factual support still needs the writerFor each finding:
Offer to edit after the audit.
Return only the corrected substantive reply.
Passes when the result is clear, specific, natural aloud, and recognizably the writer's.
Fails when it merely swaps buzzwords, erases personality, invents support, turns the piece into generic professional prose, or lets a known pattern through because the skill was treated as passive guidance.