Scan a draft for every symptom in the author's allergy list (AI-tell constructions, lexicon, fake-specific numbers, em-dash density), quote each hit, and rewrite only the flagged passages until a full re-scan is clean. Flags, never scores. Use on any draft before it goes out, as the first review lens, or when the author asks whether something "sounds like AI."
95%
Does it follow best practices?
Run evals on this skill
Adds up to 20 points to the overall score
Passed
No findings from the security scan
A veto, not a score. A draft is either clean or flagged, and a flagged draft doesn't go out. There are no thresholds and no "minor enough" exceptions.
Allergist: clean. 11 symptoms checked.
Allergist flag, A1 (the "it's not X, it's Y" reversal): "It's not about the tooling. It's about the habit."
voice skill's voice guide. Leave everything
outside the span untouched.Allergist: 2 flags.
- A2 (punchline-restatement): "…and that's exactly the problem."
- A9 (AI-tell lexicon): "leverage"
Rewrote both passages; re-scan clean.
Don't edit the list here. Hand the catch to the lessons skill, which drafts
the new entry and adds it only after the author approves.
shortrib-labs/content-loop
1e17068
If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.