24 lines
1.1 KiB
Markdown
24 lines
1.1 KiB
Markdown
---
|
|
name: codex-agent
|
|
description: Delegate execution to Codex — write code, fix bugs, research. Use when user says "delegate to codex", "have codex do", or a plan step is assigned to Codex.
|
|
---
|
|
|
|
# Delegating a task to Codex
|
|
|
|
Spawn the `codex` subagent. See `../references/common.md` for transport details (path relative to SKILL.md).
|
|
|
|
## Subagent prompt — required
|
|
|
|
- `You are a wrapper around the codex CLI: do not solve the task yourself, launch codex and return its answer + thread_id.`
|
|
- `cwd: <absolute project path>`
|
|
- Task in `=== TASK FOR CODEX ===` / `=== END TASK ===` with acceptance criteria + what to return.
|
|
- Optional `codex options: -m <model> -s workspace-write ...` (model/effort from discovery, not hardcoded).
|
|
|
|
## Discovery
|
|
|
|
If user asks which Codex models/reasoning are available, run `scripts/codex-models.sh` or `codex debug models | jq '[.models[] | select(.visibility=="list") | {slug, levels: [.supported_reasoning_levels[].effort]}]'` and present table. Default model/reasoning = `~/.codex/config.toml`.
|
|
|
|
## After
|
|
|
|
Verify diff + tests yourself. Refinements via SendMessage to same agent.
|