[kyle-sexton]

Refs: melodic-software/claude-code-plugins#13586

Summary

Before this change, an empty subscription token in either reusable stopped the lane, even when FALLBACK_CLAUDE_CODE_OAUTH_TOKEN was set. Now both reusables hand that case to the fallback leg, the same rule an empty key already follows to the subscription leg. The owner asked for this as part of making the lanes "consistent and uniform" (#13586). The claude-code-plugins lanes already follow it (#13590, #13608).

Fix

• intake-triage.yml:
  • Each leg now outputs fallback-ready.
  • The fallback job starts on handoff == 'true', or on credential-present == 'false' && fallback-ready == 'true'.
  • Its lane record names fallback reason no-key.
• pr-review-cli.yml: the same change for review-fallback.
• With every token empty, nothing new starts, and the existing status and record logic report it as before.

Verification

• node --test .github/scripts/*.test.cjs: 912 tests, 912 pass.
• New assertions in intake-triage-legs.test.cjs and pr-review-cli.test.cjs. An empty subscription token with the fallback token set starts the fallback leg. Without the fallback token, or with no credential check at all, it does not. The leg records no-key.
• zizmor --offline: no findings. actionlint reports only the existing $/ self-reference format.

Related

• Not in this PR: the check-key-credit gate for intake-triage.yml. pr-review-cli.yml already has it. Every caller pins intake-triage.yml@main and grants no actions: read, so a gate job added here would break those callers the moment this merges. The callers have to add actions: read first.
• claude-code-plugins#13608: the same rule for that repository's lanes.

🤖 Generated with Claude Code

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-11-05T21:47:01.279845Z	899c0ec	Draft marked ready

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[github-actions]

claude-review reviewed 422abd9f625bbdb38fc79944992f9dc26f0cc7e5: 0 finding(s), 0 as inline comments.

Incremental review of 422abd9, covering .github/scripts/intake-triage-legs.test.cjs and .github/scripts/pr-review-cli.test.cjs. I found nothing to report.

The change only reflows long assert.equal(...) calls onto multiple lines in the way the repository's biome formatter does. I compared each reflowed call with its original in .claude-lane/incremental.diff, and the inputs and expected values are the same:

• intake-triage-legs.test.cjs has two starts(...) assertions (expecting true and false) and one evaluate(...) assertion (expecting "no-key").
• pr-review-cli.test.cjs has three starts(...) assertions (expecting true, false, false). The only other difference is that a stray space before a comma is gone.

I did not run the tests; this verdict comes from reading the diff. I also did not check the formatting against biome. The PR context, settled-threads and prior-findings files were not needed for a formatting-only change.

[github-actions]

claude-security-review reviewed 422abd9f625bbdb38fc79944992f9dc26f0cc7e5: 0 finding(s), 0 as inline comments.

Incremental review of 422abd9, covering the changes since 899c0ec in .github/scripts/intake-triage-legs.test.cjs and .github/scripts/pr-review-cli.test.cjs.

No security findings.

Both changes only reformat existing assert.equal(...) calls across several lines. They add, remove, or change no assertion, input value, expected result, or code path. The files are test-only, and neither is a trust control, trusted path, plugin manifest, or instruction surface.

critical-findings: 0