Codex Review

OpenCode Plan Review (Iterative)

Send the current implementation plan to OpenCode (using GPT-5.6-Sol) for review. Claude revises the plan based on feedback and re-submits until approved. Max 5 rounds.

When to Invoke

• When the user runs /codex-review during or after plan mode
• When the user wants a second opinion on a plan from a different model

Agent Instructions

When invoked, perform the following iterative review loop:

Step 1: Generate Session ID

Generate a unique ID to avoid conflicts with other concurrent Claude Code sessions:

REVIEW_ID=$(uuidgen | tr '[:upper:]' '[:lower:]' | head -c 8)

Use this for all temp file paths: /tmp/claude-plan-${REVIEW_ID}.md and /tmp/opencode-review-${REVIEW_ID}.md.

Step 2: Capture the Plan and Repo Context

Write the current plan to the session-scoped temporary file. The plan is whatever implementation plan exists in the current conversation context (from plan mode, or a plan discussed in chat).

1. Detect the repo context — run git remote get-url origin 2>/dev/null to get the repo URL. If working in a subdirectory, note the relative path.
2. Write the plan file to /tmp/claude-plan-${REVIEW_ID}.md with repo context at the top:

## Repository Context

- **Repo:** https://github.com/org/repo (or local path if no remote)
- **Working directory:** /path/to/current/dir
- **Branch:** current-branch-name

---

## Implementation Plan

[The actual plan content here]

3. If there is no plan in the current context, ask the user what they want reviewed

Step 3: Initial Review (Round 1)

Run OpenCode CLI in non-interactive mode to review the plan:

opencode run \
  -m openai/gpt-5.6-sol \
  "Review the implementation plan in /tmp/claude-plan-${REVIEW_ID}.md.

The plan includes repository context at the top — use this to understand the codebase. You can browse the repo for additional context if needed.

Focus on:
1. Correctness - Will this plan achieve the stated goals?
2. Risks - What could go wrong? Edge cases? Data loss?
3. Missing steps - Is anything forgotten?
4. Alternatives - Is there a simpler or better approach?
5. Security - Any security concerns?
6. Codebase fit - Does this align with existing patterns in the repo?

Be specific and actionable. If the plan is solid and ready to implement, end your review with exactly: VERDICT: APPROVED

If changes are needed, end with exactly: VERDICT: REVISE" > /tmp/opencode-review-${REVIEW_ID}.md 2>&1

Capture the OpenCode session ID from the output. OpenCode displays session info that can be used with --session to continue. Store this as OPENCODE_SESSION_ID.

Notes:

• Use -m openai/gpt-5.6-sol as the default model. If the user specifies a different model (e.g., /codex-review anthropic/claude-sonnet-5), use that instead.
• OpenCode has read access to the filesystem by default for reading context.
• Redirect stdout to capture the output to a file for reliable reading.

Step 4: Read Review & Check Verdict

1. Read /tmp/opencode-review-${REVIEW_ID}.md
2. Present OpenCode's review to the user:

## OpenCode Review — Round N (model: openai/gpt-5.6-sol)

[OpenCode's feedback here]

3. Check the verdict:
   • If VERDICT: APPROVED → go to Step 7 (Done)
   • If VERDICT: REVISE → go to Step 5 (Revise & Re-submit)
   • If no clear verdict but feedback is all positive / no actionable items → treat as approved
   • If max rounds (5) reached → go to Step 7 with a note that max rounds hit

Step 5: Revise the Plan

Based on OpenCode's feedback:

1. Revise the plan — address each issue OpenCode raised. Update the plan content in the conversation context and rewrite /tmp/claude-plan-${REVIEW_ID}.md with the revised version.
2. Briefly summarize what you changed for the user:

### Revisions (Round N)
- [What was changed and why, one bullet per issue addressed]

3. Inform the user what's happening: "Sending revised plan back to OpenCode for re-review..."

Step 6: Re-submit to OpenCode (Rounds 2-5)

Continue the existing OpenCode session so it has full context of the prior review:

opencode run \
  -m openai/gpt-5.6-sol \
  --session ${OPENCODE_SESSION_ID} \
  "I've revised the plan based on your feedback. The updated plan is in /tmp/claude-plan-${REVIEW_ID}.md.

Here's what I changed:
[List the specific changes made]

Please re-review. If the plan is now solid and ready to implement, end with: VERDICT: APPROVED
If more changes are needed, end with: VERDICT: REVISE" > /tmp/opencode-review-${REVIEW_ID}.md 2>&1

Note: If session continuation fails (e.g., session expired), fall back to a fresh opencode run call with context about the prior rounds included in the prompt.

Then go back to Step 4 (Read Review & Check Verdict).

Step 7: Present Final Result

Once approved (or max rounds reached):

## OpenCode Review — Final (model: openai/gpt-5.6-sol)

**Status:** ✅ Approved after N round(s)

[Final feedback / approval message]

---
**The plan has been reviewed and approved. Ready for your approval to implement.**

If max rounds were reached without approval:

## OpenCode Review — Final (model: openai/gpt-5.6-sol)

**Status:** ⚠️ Max rounds (5) reached — not fully approved

**Remaining concerns:**
[List unresolved issues from last review]

---
**OpenCode still has concerns. Review the remaining items and decide whether to proceed or continue refining.**

Step 8: Cleanup

Remove the session-scoped temporary files:

rm -f /tmp/claude-plan-${REVIEW_ID}.md /tmp/opencode-review-${REVIEW_ID}.md

Loop Summary

Round 1: Claude sends plan → OpenCode reviews → REVISE?
Round 2: Claude revises → OpenCode re-reviews (session continue) → REVISE?
Round 3: Claude revises → OpenCode re-reviews (session continue) → APPROVED ✅

Max 5 rounds. Each round preserves OpenCode's conversation context via session continuation.

Rules

• Claude actively revises the plan based on OpenCode feedback between rounds — this is NOT just passing messages, Claude should make real improvements
• Always include repo context — detect the git remote URL and working directory so OpenCode can browse the codebase for context
• Default model is openai/gpt-5.6-sol. Accept model override from the user's arguments (e.g., /codex-review anthropic/claude-sonnet-5)
• Max 5 review rounds to prevent infinite loops
• Show the user each round's feedback and revisions so they can follow along
• If OpenCode CLI is not installed or fails, inform the user and suggest installing from https://opencode.ai or running curl -fsSL https://opencode.ai/install | bash
• If a revision contradicts the user's explicit requirements, skip that revision and note it for the user