Opus finding
`OPERATOR_AUTHORED` derived from comparing enabled-skill names to parent `skills/` directory listing is unreliable
lowbugmedium
- skills/contributor-spotlight/SKILL.md:163-168
`PARENT_REPO` is referenced but never defined earlier in the spec. Also `gh api .../contents/skills` paginates at 1000 entries; the parent has many skills (visible in `aeon.yml`). Listing parent skills only by directory name compared to enabled keys may produce false positives (e.g. skill renamed upstream) or false negatives (skill imported with same name from a third party). The newsworthy claim "operator-authored or operator-imported novel skills" depends on this comparison being correct; getting it wrong produces public recognition posts that misattribute authorship — a reputational/social bug worth fixing.
Recommendation
Define `PARENT_REPO` explicitly (probably from `git config --get remote.upstream.url` or a known constant). Cross-check operator-authored skills by reading the fork's `skills/<name>/SKILL.md` and confirming it does not exist upstream (or hashes differ). Note in the article that the marking is heuristic.