codex-opinion

Codex Second Opinion

Get a second opinion from OpenAI Codex on your current work. Codex uses your configured model and settings.

How to call

Two options — pick whichever fits:

Option A: Foreground (Bash)

Use when you need Codex's response before proceeding.

echo "<gathered context>" | python3 ${CLAUDE_PLUGIN_ROOT}/skills/codex-opinion/scripts/ask_codex.py

Option B: Background (Monitor)

Use when you want to keep working while Codex analyzes. Run the same command directly via Monitor — do NOT use TaskCreate, do NOT monitor a file, do NOT sleep+cat. Monitor runs the command itself and notifies you when output arrives.

echo "<gathered context>" | python3 ${CLAUDE_PLUGIN_ROOT}/skills/codex-opinion/scripts/ask_codex.py

Set Monitor timeout to at least 600 seconds. Codex can take several minutes.

Building the context

If there are uncommitted changes:

git diff HEAD | python3 ${CLAUDE_PLUGIN_ROOT}/skills/codex-opinion/scripts/ask_codex.py

If there are also untracked files that matter, include them:

{ git diff HEAD; echo "--- Untracked files ---"; git ls-files --others --exclude-standard | while read f; do echo "=== $f ==="; cat "$f"; done; } | python3 ${CLAUDE_PLUGIN_ROOT}/skills/codex-opinion/scripts/ask_codex.py

If the tree is clean, gather context yourself:

echo "Review this codebase. Key files: ..." | python3 ${CLAUDE_PLUGIN_ROOT}/skills/codex-opinion/scripts/ask_codex.py

With a custom instruction (from user text after /codex-opinion:codex-opinion):

echo "<context>" | python3 ${CLAUDE_PLUGIN_ROOT}/skills/codex-opinion/scripts/ask_codex.py "user's instruction"

Never pipe an empty string. If git diff HEAD would be empty, use echo with gathered context instead.

Session continuity

One Codex session per project, persisted across Claude Code sessions. Follow-up calls resume the prior Codex thread so it keeps its accumulated codebase knowledge. If the session has expired server-side, the script logs a notice and starts fresh.

After Codex responds

Tell the user what Codex found.