Repository navigation
Dev: Add REVIEW.md with review priorities for Claude Code Review - #1041
Merged
Merged
Conversation
Tells the Code Review service what failure looks like in OpenStation, what counts as Important, which rules to check on every pull request, what to skip, and how to keep review rounds converging. The rules point at their reasoning in AGENTS.md instead of restating it.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What it does
Adds a root
REVIEW.md, the file Claude Code Review reads to decide what to flag on a pull request, at what severity, and what to skip. It changes nothing until the Code Review service is enabled for this repository.Rationale
AGENTS.mdsays what the rules are. It doesn't say which broken rule should block a merge and which is a nit, so a reviewer weighs every rule the same. Most PRs here are written by agents, so each nit costs another commit and another review round. The file applies the guidance from Anthropic's write-up of their own review setup: keep it short, say what failure looks like in this repo, cap the nits, list paths to skip, require evidence, and fix the summary format.Implementation
AGENTS.mdsection instead of restating it. The list includes the read-path gates, frozendesktop_mode_*names,assets/js/build output, the Legacy snapshot, cross-bundle state, settings that need a menu refresh, docs that no longer match the code, and the palette rules for feature stylesheets (which no test sweeps).languages/, lockfiles, missing test coverage, and file-length warnings. A higher bar applies inbin/,tools/,reports/,extensions/andtests/.file:linefor every finding, at most three Nits, one finding per bug class, no repeated findings on a re-review, and a summary that opens with the counts.How the service reads the file:
REVIEW.mdfrom trunk, never from the PR's own changes, so edits to this file apply only after it merges. Because trunk is protected, the instructions that lower a severity, skip paths or cap nits are all honored.@-includeAGENTS.md.@claudeworkflow.claude.ymlruns Claude Code itself, which loadsCLAUDE.mdbut notREVIEW.md. Neither does the local/code-reviewcommand.bin/package.shis an allow-list, so the file stays out of the release zip.Testing instructions
Nothing to run locally. Once Code Review is enabled for the repository, comment
@claude reviewon an open PR. Each finding should carry a 🔴/🟡/🟣 label, the review should post at most three Nits, and the summary should open with a line like1 Important, 2 Nit, 0 Pre-existing.