1---2name: second-opinion3description: Ask another local AI CLI for feedback on a proposal, architecture, feature plan, refactor, review, or technical decision, then compare the feedback and produce a revised assessment. Use when the user wants another model's opinion or asks to check a plan with Gemini, Claude, Codex, Cursor, or another installed AI tool.4---56# Second Opinion78## Input910- A proposal, architecture, feature plan, refactor plan, implementation approach, review finding, or technical decision to sanity-check.11- Use explicit input first; otherwise infer the proposal and target AI from context.12- Safest default: detect available local AI CLIs and ask which one to use.1314## Workflow15161. **Clarify target and input**. Identify the proposal to review and the AI CLI to ask. If either is missing, infer from context or ask one focused question.172. **Detect available CLIs**. Check likely commands with `command -v`, such as `claude`, `gemini`, `codex`, `cursor`, `opencode`, and `aider`. Do not assume a CLI exists.183. **Verify invocation**. Use the selected CLI's help output before running it. Prefer non-interactive prompt modes and avoid commands that open editors, mutate files, install packages, or start long-lived sessions.194. **Write the proposal**. Create a temporary directory with `mktemp -d` and write the proposal, relevant constraints, and explicit questions to a file such as `proposal.md`.205. **Ask the other AI**. Pass the proposal file content to the selected CLI. Ask for critique, missed risks, alternatives, and concrete changes to the proposal.216. **Capture feedback**. Save the raw response in the same temporary directory. If the CLI fails, report the command, failure mode, and closest safe fallback.227. **Compare judgments**. Separate useful objections from weak or irrelevant feedback. Do not accept the other AI's answer just because it came from another model.238. **Revise if warranted**. Produce a final assessment, noting what changed, what did not, and why.2425## Output2627- Target AI and CLI used28- Proposal file path and feedback file path29- Other AI's strongest points30- Points rejected or deprioritized31- Final assessment32- Revised proposal when useful3334## Examples3536- `second-opinion with gemini on this architecture plan`37- `ask Claude for a second opinion on the refactor`38- `get another model to critique this feature proposal`39- `compare our deploy plan with Codex`4041## Guardrails4243- Do not send secrets, credentials, private keys, tokens, customer data, or unnecessary proprietary context to another CLI.44- Do not run a paid, networked, or external AI CLI unless the user requested that tool or approved it.45- Do not let the other AI mutate the repository; use it for feedback only.46- Keep the prompt focused. Send the proposal and constraints, not the entire repo, unless explicitly requested.47- Treat second opinions as advisory evidence, not authority.
hunvreus