Opus finding
WoW release delta is misleading because the prior snapshot is also a 7-day window measured at run time, not a disjoint prior week
mediumbughigh
- skills/fleet-state/SKILL.md:134-140
- skills/fleet-state/SKILL.md:92-100
Both the current and prior `release_count` are computed as "announced entries with `published_at` ≥ now-7d at the moment of the run". Two consecutive Monday runs both see the trailing 7-day window relative to *their* run date, which overlaps by 0 days only if runs are exactly 7 days apart. In practice an off-by-a-few-hours run will double-count releases that fall in the overlap, making `delta_releases` a near-meaningless metric and falsifying the inline comment claiming it is real WoW velocity. Combined with the fact that the verdict line `SHIPPED: ${N} fork releases this week` doesn't reference the delta, this is more about a deceptive comment + a misleading WoW row in the article. Misleading deltas in an operator-facing digest erode trust per the skill's own "Quality bar".
Recommendation
Either (1) compute delta_releases against the count of releases whose `published_at` falls strictly *between* prior_run_ts and now, or (2) drop delta_releases as a WoW metric and document that release velocity is a rolling-window count, not a WoW delta. Update the comment so it does not claim WoW semantics it does not provide.