fix: pin .claude markdown to LF so the pr-review parity gate compares equal bytes - #466
fix: pin .claude markdown to LF so the pr-review parity gate compares equal bytes#466thomasluizon wants to merge 1 commit into
Conversation
… equal bytes
tools/launch-worker.mjs in orbit-ui-mobile sha256s the working-tree bytes of
.claude/skills/pr-review/{SKILL.md,rubric.md} in both repositories and refuses to
launch an independent reviewer for repo ui or api when they differ.
The committed blobs are already byte-identical:
SKILL.md b374cfd
rubric.md 308d1c7
Both repositories set core.autocrlf=true. orbit-ui-mobile pins
`.claude/skills/**/*.md text eol=lf`; this repository pinned only *.mjs and two
JSON artifacts. The same blob therefore checked out LF there and CRLF here,
+230 bytes on SKILL.md and +289 on rubric.md. `git check-attr text eol` reported
`eol: lf` in the UI repository and `eol: unspecified` here. `git status` was clean
in both, which is why nothing looked wrong.
Cost, measured on the overnight run of 2026-08-08: 76 tickets stood down rather
than opened unreviewable, and 4 pull requests handed over with no independent
review.
This commit changes no blob. It pins `.claude/**/*.md text eol=lf`, a superset of
the missing line that also covers every future mirrored document under .claude,
and re-materializes the working tree from the unchanged blobs.
Proof, both working trees after the change:
SKILL.md 3fbf6d133b4a7bafbe8da8345ea10d45ff96ac6b83d6c78f348651fa3c830059 13388 B
rubric.md 672f14194f2bc7834828d434babdad152d5b361b2059b75f34f5b0d370be72b6 20247 B
Gate proof: with .gitattributes reverted and the tree re-materialized, the launcher
prints "pr-review parity failed for .claude\skills\pr-review\SKILL.md". With this
commit applied it passes that check and advances to the next one.
Audit of every other path present in both repositories: 30 shared paths, 3 with an
identical committed blob, and only these 2 drifted. The other 27 differ in content
on purpose.
Refs ORB-215
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 5eca6803ae
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".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| # blob as CRLF here and LF there, +230 bytes on SKILL.md and +289 on rubric.md, so the | ||
| # gate compared different bytes for the same content and stood 76 tickets down on | ||
| # 2026-08-08. `git status` stayed clean in both repos, which is why nothing looked wrong. | ||
| .claude/**/*.md text eol=lf |
There was a problem hiding this comment.
When this commit is pulled into an existing core.autocrlf=true checkout whose .claude Markdown files already contain CRLF, Git updates .gitattributes but leaves these unchanged blobs untouched, so the launcher continues hashing the 13,618-byte CRLF SKILL.md and the parity gate remains blocked. Reproducing a checkout from the parent to this commit retained all 230 CR bytes despite git check-attr reporting eol: lf; this rollout therefore needs an explicit worktree refresh or the blob-based launcher change to land simultaneously.
Useful? React with 👍 / 👎.
|
1 participant
What this fixes
tools/launch-worker.mjs:99-112in orbit-ui-mobile hashes the working-tree bytes of.claude/skills/pr-review/{SKILL.md,rubric.md}in both repositories. It refuses to launch the independent reviewer forrepo uiorrepo apiwhen the two hashes differ.The committed blobs are already identical in both repositories:
Both repositories set
core.autocrlf=true. orbit-ui-mobile pins.claude/skills/**/*.md text eol=lf. This repository pinned only*.mjsand two JSON artifacts. The same blob therefore checked out as LF there and CRLF here.git statuswas clean in both repositories. That is why nothing looked wrong.Cost, measured on the overnight run of 2026-08-08: 76 tickets stood down rather than opened unreviewable, and 4 pull requests handed over with no independent review.
What changed
One line in
.gitattributes:.claude/**/*.md text eol=lf. That is a superset of the missing.claude/skills/**/*.mdline, and it also covers every future mirrored document under.claude.No blob changes. The committed content was already correct. Only the working tree was wrong, so the working tree is re-materialized from the unchanged blobs.
Proof
Both working trees after the change:
Those are the exact hashes recorded on ORB-215.
Red-capable gate proof. The guard was removed, the failure observed, then the guard restored:
With the fix, the launcher passes the parity check and stops at the next check (the branch check). Without it, the launcher prints the exact refusal that stood 76 tickets down.
Audit of other pinned pairs
Question asked on ORB-215: has any other path pair drifted the same way?
No. Full sweep of every path present in both repositories at HEAD:
After this change that count is 0. The other 27 shared paths carry different blobs on purpose.
.claude/skills/_shared/verification-protocol.mdand.claude/skills/second-opinion/SKILL.mddiffer in real content, not line endings: the API copy of the first carries a mirror header, and the second omits the UI-only unattended-run sentences.Related
A checkout setting must never be able to fail a contract gate again. The durable half of this fix is in the paired orbit-ui-mobile pull request, which makes the launcher compare committed blobs instead of working-tree bytes. This commit alone holds only on a machine whose
.gitattributesis applied; the blob comparison holds on a fresh clone too.Refs ORB-215