chore(ci): bump review-cli to 1.10.6 - #145
Conversation
Move the `REVIEW_CLI_VERSION` fallback in claude-code-review.yml from 1.10.1 to 1.10.6, and update the repository-variables table in .github/workflows/CLAUDE.md so the documented default matches. The literal is only the fallback behind the `vars.REVIEW_CLI_VERSION` repo variable, which is not set on this repo today, so the fallback is load-bearing.
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
🤖 Claude PR Metadata GenerationStatus: ❌ Error
|
|
● Reviewed · against Note Approved. Bumps the pinned review-cli version from 1.10.1 to 1.10.6 in the CI workflow fallback and matching documentation. AssessmentTwo-line change: the This is a CI-only dev tool—installed ad-hoc at job runtime, never in The only stale reference is the comment at line 108 ("Fallback matches what Uniswap/universe runs today"), which the PR description explicitly left alone to avoid widening the diff. Worth confirming Tip Teach the reviewer. React 👍 on findings that helped, 👎 on false positives. Reply to push back or add context — we aggregate this weekly to tune the bot. Comment |
There was a problem hiding this comment.
Note
✅ Approved — see full review in the sticky comment ↑
Requested by Erin Hales · Slack thread
Bumps the pinned
@uniswap/review-cliversion from1.10.1to1.10.6.What changed
Two lines, both under
.github/workflows/:claude-code-review.yml:109— theREVIEW_CLI_VERSIONenv:fallback, consumed by thebun add "@uniswap/review-cli@${REVIEW_CLI_VERSION}"step. That var is declared once at workflow level and read by both install blocks (thetriagejob at line 220 and thereviewjob at line 426), so the single edit covers both and they cannot drift apart.CLAUDE.md:329— the Repository Variables table, which states the default in prose (defaults to 1.10.1). Updated alongside the workflow; leaving it would have made the docs contradict the workflow.Nothing else changed. There is deliberately no
package.jsonentry and no lockfile regeneration:review-cliis installed ad-hoc into$RUNNER_TEMPat job time and has no manifest or lockfile reference in this repo. A repo-wide grep confirms no otherreview-cliversion literal remains.This is only the fallback — the repo variable still needs bumping
vars.REVIEW_CLI_VERSIONis the authoritative pin;1.10.6is just the||fallback behind it. A PR cannot set a repo variable, so ifvars.REVIEW_CLI_VERSIONis set on this repo it will keep winning and merging this alone changes nothing at runtime. It needs bumping separately in repo settings → Variables for this to take effect.Per
.github/workflows/CLAUDE.md, onlyNODE_VERSIONandNPM_VERSIONare actually set on this repo today, so as documented the fallback is currently load-bearing and this change should take effect on merge. Worth a quick confirm against live repo settings, since the doc could itself be stale.Validation
env.REVIEW_CLI_VERSIONresolves to the expected stringbun run lint:actions(actionlint v1.7.7)markdownlint-cli2on the edited docbunx nx format:check --uncommittedlefthookpre-commit (actionlint,format,lint)1.10.1/review-cliThe only remaining
1.10.1hits in the tree are theuniswap-tradingplugin's own package version (a coincidental collision, correctly untouched) and unrelated transitive deps inbun.lock. Noversion:key was touched — the diff is exactly two lines.bun run docs:lint(vale) reports 296 pre-existing spelling errors across 40 files, all indocs/. That is unrelated to this PR: vale is scoped todocs/**/*.mdby.vale.ini, andgit diff origin/main -- docs/is empty, so its output is byte-identical before and after this change. It also means vale does not cover the.github/workflows/CLAUDE.mdedit —markdownlint-cli2does, and it passes.Note for reviewers
The comment directly above the edited line still reads "Fallback matches what Uniswap/universe runs today." I left it alone rather than widen the diff, but it is only accurate if
universeis on1.10.6too — worth a glance if these bumps are landing at different times across repos.Generated by Claude Code