codex-claude-cursor-loop

What this skill does

# Codex-Claude-Cursor Engineering Loop Skill

## Core Workflow Philosophy
This skill implements a 3-way sequential validation engineering loop:
- **Claude Code**: Architecture and planning, final review
- **Codex**: Plan validation (logic/security), code review (bugs/performance)
- **Cursor Agent**: Code implementation and execution
- **Sequential Validation**: Claude plans → Codex validates → Cursor implements → Codex reviews → Claude final check → repeat

## Phase 1: Planning with Claude Code
1. Start by creating a detailed plan for the task
2. Break down the implementation into clear steps
3. Document assumptions and potential issues
4. Output the plan in a structured format

## Phase 2: Plan Validation with Codex
1. Ask user (via `AskUserQuestion`):
   - Model: `gpt-5` or `gpt-5-codex`
   - Reasoning effort: `low`, `medium`, or `high`
2. Send the plan to Codex for validation:
```bash
   echo "Review this implementation plan and identify any issues:
   [Claude's plan here]

   Check for:
   - Logic errors
   - Missing edge cases
   - Architecture flaws
   - Security concerns" | codex exec -m <model> --config model_reasoning_effort="<effort>" --sandbox read-only
```
3. Capture Codex's feedback and summarize to user

## Phase 3: Plan Refinement Loop
If Codex finds issues in the plan:
1. Summarize Codex's concerns to the user
2. Refine the plan based on feedback
3. Ask user (via `AskUserQuestion`): "Should I revise the plan and re-validate, or proceed with implementation?"
4. Repeat Phase 2 if needed until plan is solid

## Phase 4: Implementation with Cursor Agent
Once the plan is validated by Codex:

### Session Management
1. Ask user (via `AskUserQuestion`): "Do you want to start a new Cursor session or resume an existing one?"
   - **New session**: Start fresh
   - **Resume session**: Continue previous work

2. If resuming:
```bash
   # List available sessions
   cursor-agent ls

   # Let user select session ID
   # Store session ID for subsequent calls
```

3. Ask user (via `AskUserQuestion`): Which Cursor model to use (e.g., `composer-1`, `claude-3.5-sonnet`, `gpt-4o`)

### Implementation
4. Send the validated plan to Cursor Agent:

**For new session:**
```bash
   cursor-agent --model "<model-name>" -p --force "Implement this plan:
   [Validated plan here]

   Please implement the code following these specifications exactly."
```

**For resumed session:**
```bash
   cursor-agent --resume="<session-id>" -p --force "Continue implementation:
   [Validated plan here]"
```

5. **IMPORTANT**: Store the session ID from the output for all subsequent Cursor calls
6. Capture what was implemented and which files were modified

## Phase 5: Codex Code Review
After Cursor implements:
1. Send Cursor's implementation to Codex for code review:
```bash
   echo "Review this implementation for:
   - Bugs and logic errors
   - Performance issues
   - Security vulnerabilities
   - Best practices violations
   - Code quality concerns

   Files modified: [list of files]
   Implementation summary: [what Cursor did]" | codex exec --sandbox read-only
```
2. Capture Codex's code review feedback
3. Summarize findings to user

## Phase 6: Claude's Final Review
After Codex code review:
1. Claude reads the implemented code using Read tool
2. Claude analyzes both:
   - Codex's review findings
   - The actual implementation
3. Claude provides final assessment:
   - Verify if it matches the original plan
   - Confirm Codex's findings are valid
   - Identify any additional concerns
   - Make final architectural decisions
4. Summarize overall quality and readiness

## Phase 7: Iterative Improvement Loop
If issues are found (by Codex or Claude):
1. Claude creates a detailed fix plan based on:
   - Codex's code review findings
   - Claude's final review insights
2. Send the fix plan to Cursor Agent using the **same session**:
```bash
   # IMPORTANT: Use --resume with the stored session ID
   cursor-agent --resume="<session-id>" -p --force "Fix these issues:
   [Detailed fix plan]

   Issues from Codex: [list]
   Issues from Claude: [list]"
```
3. After Cursor fixes, repeat from Phase 5 (Codex code review)
4. Continue the loop until all validations pass
5. **Note**:
   - Use same Codex model for consistency
   - Always use the same Cursor session ID to maintain context
   - Session maintains full history of changes

## Recovery When Issues Are Found

### When Codex finds plan issues (Phase 2):
1. Claude analyzes Codex's concerns
2. Refines the plan addressing all issues
3. Re-submits to Codex for validation
4. Repeats until Codex approves

### When Codex finds code issues (Phase 5):
1. Claude reviews Codex's findings
2. Creates detailed fix plan
3. Sends to Cursor for fixes
4. After Cursor fixes, back to Codex review
5. Repeats until Codex approves

### When Claude finds issues (Phase 6):
1. Claude creates comprehensive fix plan
2. Sends to Cursor for implementation
3. After fixes, Codex reviews again
4. Claude does final check
5. Repeats until Claude approves

## Best Practices
- **Always validate plans with Codex** before implementation
- **Never skip Codex code review** after Cursor implements
- **Never skip Claude's final review** for architectural oversight
- **Maintain clear handoff** between all three AIs
- **Document who did what** for context
- **Use same models** throughout (same Codex model, same Cursor model)
- **Session Management**:
  - Always use `--resume` with same session ID for iterative fixes
  - Store session ID at the start and reuse throughout
  - Use `cursor-agent ls` to find previous sessions
  - Only start new session when beginning completely new feature

## Command Reference
| Phase | Who | Command Pattern | Purpose |
|-------|-----|----------------|---------|
| 1. Plan | Claude | TodoWrite, Read, analysis tools | Claude creates detailed plan |
| 2. Validate plan | Codex | `echo "plan" \| codex exec -m <model> --config model_reasoning_effort="<effort>" --sandbox read-only` | Codex validates logic/security |
| 3. Refine | Claude | Analyze Codex feedback, update plan | Claude fixes plan issues |
| 4. Session setup | Claude + User | Ask new/resume, `cursor-agent ls` if needed | Setup or resume Cursor session |
| 5. Implement | Cursor | `cursor-agent --model "<model>" -p --force "prompt"` OR `cursor-agent --resume="<id>" -p --force "prompt"` | Cursor implements validated plan |
| 6. Review code | Codex | `echo "review" \| codex exec --sandbox read-only` | Codex reviews for bugs/performance |
| 7. Final review | Claude | Read tool, analysis | Claude final architectural check |
| 8. Fix plan | Claude | Create detailed fix plan | Claude plans fixes from all feedback |
| 9. Apply fixes | Cursor | `cursor-agent --resume="<id>" -p --force "fixes"` | Cursor implements fixes in same session |
| 10. Re-review | Codex + Claude | Repeat phases 6-7 | Validate fixes until perfect |

## Error Handling
1. Monitor Cursor Agent output for errors
2. Summarize Cursor's implementation results and Claude's review
3. Ask for user direction via `AskUserQuestion` if:
   - Significant architectural changes needed
   - Multiple files will be affected
   - Breaking changes are required
4. When issues appear, Claude creates a detailed fix plan before sending to Cursor

## The Perfect Loop
```
1. Plan (Claude)
   ↓
2. Validate Plan (Codex) → if issues → refine plan → repeat
   ↓
3. Implement (Cursor)
   ↓
4. Code Review (Codex) → captures bugs/performance issues
   ↓
5. Final Review (Claude) → architectural check
   ↓
6. Issues found? → Fix Plan (Claude) → Implement Fixes (Cursor) → back to step 4
   ↓
7. All passed? → Done! ✅
```

This creates a triple-validation, self-correcting, high-quality engineering system where:
- **Claude**: All planning, architecture, and final oversight
- **Codex**: All validation (plan logic + code quality)
- **Cursor Agent**: All implementation and coding