Go review
Supplement code review with Go-specific correctness and lifecycle analysis. Use alongside code-review whenever the requested scope contains meaningful Go production or test code, a Go package, or a full Go codebase; skip only when no meaningful Go code is in scope.From its SKILL.md
npx -y skills add yarlson/yarstack --skill go-reviewAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
2 things to look at
- 3 stars3 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
- runs commandsInstructs the agent to run 1 command, including `go test -race`.
SKILL.md
1.7 KB, 283 tokens by cl100k_base, as published. Nobody here has run it
Go Review
Run the required Go-specific pass alongside general review and find runtime failures it is unlikely to catch. Keep the review read-only.
Workflow
- Confirm the scope and Go version, then read relevant entry points, callers, tests, and shutdown wiring.
- Trace goroutine ownership, cancellation, error propagation, joining, blocked operations, and fan-out bounds.
- Check channel close ownership, lock ordering, atomics, concurrent collection access, contexts, timeouts, retries, tickers, and timers.
- Check cleanup of files, bodies, rows, transactions, subscriptions, and temporary resources on every relevant path.
- Check Go-specific nil, assertion, zero-value, slice-aliasing, loop-capture, panic, HTTP, database, and version-sensitive semantics only where the changed flow uses them.
- Run trusted focused tests or
go test -raceonly when authorized and useful for a concrete risk.
Use code-review for general diff ownership, security-review for trust boundaries, test-gap-review for comprehensive coverage, and rollout-readiness-review for deployment readiness. Their invocation does not authorize edits or unsafe execution.
Report affected location, runtime scenario, impact, smallest safe correction, verification, and uncertainty. State directly when no Go-specific finding exists.
Finish when the selected Go lifecycle surface is exhausted and evidence limits are explicit.
What ships with it
Read from the repository
Just SKILL.md. No reference files, no scripts.