second-opinion

Second Opinion

Shell out to external LLM CLIs for an independent code review powered by a separate model. Supports OpenAI Codex CLI (codex) and Google Antigravity CLI (agy).

When to Use

• Getting a second opinion on code changes from a different model
• Reviewing branch diffs before opening a PR
• Checking uncommitted work for issues before committing
• Running a focused review (security, performance, error handling)
• Comparing review output from multiple models

When NOT to Use

• Neither Codex CLI nor Antigravity CLI is installed
• No API key or subscription configured for either tool
• Reviewing non-code files (documentation, config)
• You want Claude's own review (just ask Claude directly)

Gemini CLI Is End-of-Life

As of 2026-06-18, gemini (Gemini CLI) stopped serving Google AI Pro, Ultra, and free-tier individual accounts. Those accounts now get UNSUPPORTED_CLIENT on every request, pointing at https://antigravity.google. The replacement is Antigravity CLI, binary agy.

Default to agy. Only fall back to gemini when the user has a Gemini Code Assist Standard/Enterprise license or a paid Gemini API key (GEMINI_API_KEY) — those still work. See references/gemini-invocation.md for that legacy path.

Quick Reference

# Codex (headless exec with structured JSON output)
codex exec -c model='"gpt-5.6-sol"' -c model_reasoning_effort='"xhigh"' \
  --sandbox read-only --ephemeral \
  --output-schema codex-review-schema.json \
  -o "$output_file" - < "$prompt_file"

# Antigravity (headless print mode — prompt must be an ARGUMENT, not stdin)
git diff HEAD > /tmp/review-diff.txt
agy --model gemini-3.1-pro-high --output-format text \
  --disable-slash-commands -p="$(cat /tmp/review-prompt.txt)"

Invocation

1. Gather context interactively

Use AskUserQuestion to collect review parameters in one shot. Adapt the questions based on what the user already provided in their invocation (skip questions they already answered).

Combine all applicable questions into a single AskUserQuestion call (max 4 questions).

Question 1 — Tool (skip if user already specified):

If the user says "gemini", treat it as Antigravity unless they explicitly have a paid Code Assist / API-key setup.

Question 2 — Scope (skip if user already specified):

Question 3 — Project context (skip if neither CLAUDE.md nor AGENTS.md exists):

Check for CLAUDE.md first, then AGENTS.md in the repo root. Only show this question if at least one exists.

Question 4 — Review focus (always ask):

2. Run the tool directly

Do not pre-check tool availability. Run the selected tool immediately. If the command fails with "command not found" or an extension is missing, report the install command from the Error Handling table below and skip that tool (if "Both" was selected, run only the available one).

agy installs to ~/.local/bin, which is not always on PATH. Prefix the call with export PATH="$HOME/.local/bin:$PATH" before concluding it is missing.

Diff Preview

After collecting answers, show the diff stats:

If the diff is empty, stop and tell the user.

If the diff is very large (>2000 lines changed), warn the user and ask whether to proceed or narrow the scope.

Auto-detect Default Branch

For branch diff scope, detect the default branch name:

Codex Invocation

See references/codex-invocation.md for full details on command syntax, prompt assembly, and the structured output schema.

Summary:

• Uses codex exec (not codex exec review) for headless operation. codex exec review has native --uncommitted / --base / --commit scope flags, but they are mutually exclusive with a custom [PROMPT], so it cannot carry project context or focus instructions. Keep the manual prompt-assembly approach.
• Model: gpt-5.6-sol, reasoning: xhigh
• Uses OpenAI's published code review prompt (fine-tuned into the model)
• Diff is generated manually and piped via stdin with the prompt
• --output-schema produces structured JSON findings
• -o captures only the final message (no thinking/exec noise)
• All three scopes (uncommitted, branch, commit) support project context and focus instructions (no limitations)
• Falls back to gpt-5.6, then gpt-5.4, on auth errors
• Output is clean JSON — parse and present findings by priority
• Set timeout: 600000 on the Bash call

Antigravity Invocation

See references/antigravity-invocation.md for full details on flags, scope mapping, and model selection.

Summary:

• Model: gemini-3.1-pro-high (effort is baked into the model slug)
• The prompt must be a command-line argument, not stdin. agy print mode does not read prompts from stdin unless --input-format stream-json is used. Assemble the prompt into a file, then pass -p="$(cat "$prompt_file")".
• Use -p=<value> with an equals sign. Bare -p swallows the next flag as its prompt value and silently discards the real one.
• Add --disable-slash-commands so a line in the untrusted diff cannot be expanded as a slash command
• Do NOT pass --mode plan. It intermittently returns a "plan artifact / click Proceed" stub instead of the review. Omit --mode entirely
• Do NOT pass --dangerously-skip-permissions; it is unnecessary for a text-only diff review and will be blocked in restricted environments
• Prefer --output-format text. --json-schema is advisory, not enforced: it emits schema-shaped JSON nested inside a response string, duplicates the payload, and ran ~3x slower in testing
• Set timeout: 600000 on the Bash call

Scope mapping for git diff (Antigravity has no built-in scope flags):

Running Both

When the user picks "Both" (the default):

1. Run Codex and Antigravity in parallel — issue both Bash tool calls in a single response. Both commands are read-only (they review diffs via external APIs) so there is no shared state or git lock contention.
2. Collect both results, then present with clear headers:

Summarize where the two reviews agree and differ.

Error Handling

Examples

Both tools (default):

Codex only with inline args:

Antigravity only:

Large diff warning: