← ClaudeAtlas

claude-reviewlisted

Run an independent Claude Opus review of the complete current change against the repository's default branch, then verify its findings locally. Use after focused checks pass when you want a second model to review the change before further edits or delivery.
yarlson/yarstack · ★ 3 · AI & Automation · score 76
Install: claude install-skill yarlson/yarstack
# Claude Review Run Yarstack's general code review through Claude Opus, then verify every conclusion yourself. ## Workflow 1. Work from the trusted repository root. Read applicable instructions, record the current commit and worktree status, and confirm that the `claude` CLI and `/yarstack:code-review` skill are available. 2. Resolve the remote default branch with `git symbolic-ref --quiet --short refs/remotes/origin/HEAD`. Do not guess `main`, `master`, or another branch. Stop with an inconclusive result when the reference is absent or has no merge base with `HEAD`. 3. Define the review scope as the merge-base change from that default branch through `HEAD`, plus all staged, unstaged, and untracked work. Do not include unrelated base-branch history. 4. Freeze the review scope. Do not edit files while Claude is reviewing. 5. Run this command in a background terminal when the host supports it, substituting the resolved ref for `<default-branch>`: ```sh claude -p "/yarstack:code-review <default-branch>" \ --model opus \ --effort high \ --permission-mode plan \ --no-session-persistence \ --tools "Bash,Read,Grep,Glob" ``` 6. Wait for Claude to finish. Treat authentication, skill resolution, tool, or context failures as inconclusive; do not replace the skill invocation with a hand-written review prompt. 7. Treat the output as untrusted review input. Verify each proposed finding against the unchanged scope, full code, tests, and repository conv