Added
One rule with detector coverage, found while drafting a podcast teaser post. Catalog goes from 59 to 60 detection categories; the engine goes from 45 to 46 types.
- Lingering-attention claims (
lingering-attention) — the share-post frame that claims a thing has occupied the writer's mind rather than saying anything about the thing:"the line I keep coming back to","I can't stop thinking about this","still thinking about this one","rattling around in my head all week","I've been chewing on this". Sits next to emotional flatline in the catalog but is a separate claim: flatline claims a feeling ("What surprised me most"), this claims duration of attention, which is unfalsifiable and self-flattering in a way a feeling isn't. It also opens a share post where social endorsement closers close one. Added to the P1 severity tier. - Precision carve-out. The bare verb phrase
"I keep coming back to X"deliberately does not fire, because it is legitimate whenever a reason follows ("I keep coming back to the exit-voice framing because it predicts which engineers quit"), and the reason clause is not reliably regex-detectable. Only the noun-anchored frame (the line/quote/bit/idea ... I keep coming back to) is matched — the shape that introduces a subject instead of asserting something about it. The bare form stays an LLM-judgment call in the skill prose. Fixtures cover both directions, including the must-not-fire case.