[modusensus]

Summary

• fix(workflows): the ClawHub publish step walked the two skill editions as flat statements
  under bash -e, so when the registry refused the second one as a duplicate, the step ended
  before the second edition was submitted and the verify step — the only thing in the job that
  could see a missing half — never ran. It is one loop now: a duplicate refusal is recorded as
  already registered and the loop continues; a genuine failure still reddens the step at the
  end, and the verify step stays the only judge of whether a listing arrived.
• docs(agents): the 0.8.0 release had no record at all (published 2026-10-17, recorded
  2026-10-21), so both 0.8.0 and 0.9.0 are written from outside readings — npm dist-tag, shasum,
  unpacked size, attestations status, release author, and a temp-prefix install that actually
  runs the published binary. Plus the ### ClawHub registry measurement of this release's
  refusal and two new pitfalls rows (the duplicate refusal; a global npm shim with no package
  behind it).

Verification

• The changed region is driven, not eyeballed: .scratch/extract-clawhub-step.py slices the
  publish region out of publish.yml itself, and a stub clawhub answers per scenario. 9
  assertions pass — including the control that runs the old text against the same refusal and
  reproduces the defect (step exits non-zero, euthyna-en never attempted), and the equivalence
  check that the clean-run argv transcript is byte-identical to the old step, so the
  refactor moves control flow only, never what gets written to the registry.
• publish.yml parses (yaml.safe_load): three jobs, and the clawhub step names unchanged.
• node --test: 324 tests, 323 pass, 0 fail, 1 skip.
• Registry state read after the fact: both slugs pass|clean|0.9.0, so the release landed on all
  three channels (npm latest: 0.9.0, GitHub release euthyna 0.9.0, ClawHub both editions).

Limit of this change

clawhub runs only on a v* tag push, so nothing here is exercised until the next release cuts
one. The next rehearsal (gh workflow run publish.yml --ref main -f publish=false) still skips
the job — it is gated on a push event or inputs.publish == true. The stub harness is therefore
the only pre-release check this path gets; the durable one is the verify step, which fails the
job when a listing does not reach decision=pass for the version and latest.

[coderabbitai]

Warning

Review limit reached

You've used all free OSS reviews for now. Wait for the free limit to reset to keep reviewing this public repository.

Next included review available in 2 minutes.

Check out review usage here.

View limit details

Limit details: You’ve used the included review currently available.

Learn how review limits work.

Review configuration:

⚙️ Run configuration

• Configuration used: defaults
• Review profile: CHILL
• Plan: Advanced
• Run ID: 89c87b03-25e1-4ab5-8ed4-fea30eaa26fa

📥 Commits

Reviewing files that changed from the base of the PR and between af4aa79 and d490cbb.

📒 Files selected for processing: 2

• .github/workflows/publish.yml
• AGENTS.md

• Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}