Execute an implementation plan by dispatching Codex-backed implementer and reviewer agents while Claude keeps spec-compliance control. Codex-backed variant.
How this skill is triggered — by the user, by Claude, or both
Slash command
/superpowers-cc-to-codex:subagent-driven-development-codexThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
<!--
Execute plan by calling codex_implement per task via the MCP tool, with two-stage review after each: Claude-side spec compliance first, then codex_review for code quality.
Core principle: Codex implementer per task + two-stage review (spec then quality) = high quality, fast iteration
You MUST dispatch codex_implement for every plan task (Step 3). You cannot implement code yourself, even for simple tasks. Spec review requires seeing EXTERNAL implementation. No dispatch = no task completion.You MUST create a task for each plan task and complete them in order.
Setup:
Per task:
TASK_BASE_SHA="$(git rev-parse HEAD)"
Substitute the captured TASK_BASE_SHA value as the structured review base for this task, so the reviewer inspects only the task's delta rather than all prior tasks.
codex_implement MCP tool.
Pass the task description as the prompt:{
"tool": "codex_implement",
"arguments": {
"taskId": "task-17",
"prompt": "Implement Task 3 from docs/superpowers/plans/2026-04-03-agent-forwarding.md. Keep the work scoped to agents/codex-implementer.md and tests/prompt-contracts/execution-workflows.test.mjs.",
"workspaceRoot": "/absolute/path/to/your/repo"
}
}
spec-review-template.md; if issues remain, resume the same implementer thread with codex_resume:{
"tool": "codex_resume",
"arguments": {
"taskId": "task-17",
"sessionId": "019d4f82-58b8-72d3-9212-2e3d3fc69bcb",
"prompt": "Fix these issues:\n- The agent still lacks tools: Bash.\n- The prompt-file override is ignored in the TDD path.",
"workspaceRoot": "/absolute/path/to/your/repo"
}
}
codex_review only after spec compliance passes:{
"tool": "codex_review",
"arguments": {
"taskId": "task-17-review",
"reviewStyle": "structured",
"scope": { "kind": "base", "base": "<captured TASK_BASE_SHA value>" },
"prompt": "Review the implementation of Task 4 from docs/superpowers/plans/2026-04-03-agent-forwarding.md. Focus on regressions, incorrect routing, and missing test coverage.",
"workspaceRoot": "/absolute/path/to/your/repo"
}
}
After all tasks:
codex_review for the entire implementation with a new task ID and the final base SHA.Handle the implementer's reported status:
| Status | Action |
|---|---|
| DONE | Proceed to spec compliance review |
| DONE_WITH_CONCERNS | Read concerns; if correctness/scope — address before review; if observations — note and proceed |
| NEEDS_CONTEXT | Provide the missing context and re-dispatch |
| BLOCKED | Assess: context problem → re-dispatch with more context; too hard → escalate approach; too large → break down; plan wrong → escalate to human |
Never ignore an escalation or force the same agent to retry without changes.
Never:
If reviewer finds issues: resume the implementer thread to fix, reviewer reviews again, repeat until approved.
codex_review MCP toolnpx claudepluginhub mzored/superpowers-cc-to-codex --plugin superpowers-cc-to-codexGuides creation, editing, and verification of skills for AI coding agents using test-driven development with subagent scenarios. Use when authoring or debugging skills.