← ClaudeAtlas

rubber-duck-reviewlisted

Cross-model code review using a second agent, tool, or read-only CLI fallback. This skill should be used when someone asks to rubber duck a change, get a second opinion, or run an independent review focused on correctness, future-proofing, and design quality.
Vivswan/skills · ★ 2 · AI & Automation · score 71
Install: claude install-skill Vivswan/skills
# Rubber Duck Review Get a second opinion on code changes from a different model while you keep ownership of the task. ## When to Apply Use this skill when someone asks for: - `/rubber-duck-review` - "rubber duck my changes" - "get a second opinion on this" - "review this with another model" ## Workflow ### 1. Pick the reviewer - Prefer a dedicated review tool when one is available. - Use a reviewer that is different from the current model when possible. - If no dedicated tool exists, fall back to a read-only CLI invocation. Pick by which model *you* are, then use the matching full pattern in step 3: - If you are currently using **Claude** → **Codex reviewer** (§3a): `codex exec --json --sandbox read-only`, fallback `copilot -p --deny-tool='write' --deny-tool='shell'`. - If you are currently using **Codex or GitHub Copilot** → **Claude reviewer** (§3b): `claude -p --permission-mode plan --verbose --output-format stream-json`, fallback `copilot -p ...`. - Always pass the prompt as `"$(cat "$prompt_file")"` (a file inside the scratch tmp dir, see §3) and append `< /dev/null`; the copy-paste blocks in §3a/§3b already do both. - Never let the reviewer write files, edit code, or run unrestricted shell commands. `--sandbox read-only` (codex) / `--permission-mode plan` (claude) lets it run read-only commands (grep, `git diff`, typecheck) but blocks writes. That self-checking makes findings concrete. (Note: a read-only sandbox can block temp-dir creation, so the reviewer