Ask Codex
Delegate a task to the OpenAI Codex CLI and report the result. Codex runs as a peer model: read peer-cli-delegation.md and apply it throughout — it owns the delegation mechanics, the follow-up loop, the critical evaluation of Codex output, and the error-handling protocol, while this file supplies the CLI name, the exec command, the resume syntax, and the flag table. Do NOT use for: git/PR workflows (use the other autopilot skills) or tasks that do not involve the Codex CLI.
Input
Parse $ARGUMENTS for an optional task description and optional --model / --effort flags. Anything missing is collected in the next step.
Running a Task
- If model or effort was not supplied in
$ARGUMENTS, ask the user (via AskUserQuestion, one prompt with two questions) which model (gpt-5.5, gpt-5.4, gpt-5.4-mini, gpt-5.3-codex-spark, gpt-5.3-codex) and which reasoning effort (xhigh, high, medium, low).
- Select the sandbox mode; default to
--sandbox read-only unless edits or network access are required.
- Assemble the command with the appropriate options:
-m, --model <MODEL>--config model_reasoning_effort="<xhigh|high|medium|low>"--sandbox <read-only|workspace-write|danger-full-access>--full-auto-C, --cd <DIR>--skip-git-repo-check"your prompt here" (final positional argument)
- Always pass
--skip-git-repo-check.
- Resume: continue a prior session with
echo "your prompt here" | codex exec --skip-git-repo-check resume --last 2>/dev/null. Insert any flags between exec and resume; per the shared block's inherit rule, pass no config flags on resume unless the user explicitly requests them.
- IMPORTANT: append
2>/dev/null to every codex exec command to suppress thinking tokens (stderr). Only show stderr when the user explicitly asks to see thinking tokens or for debugging.
- IMPORTANT (stdin):
codex exec always reads stdin and concatenates it with the positional prompt — apply the shared block's stdin guard, e.g. codex exec ... "prompt" </dev/null 2>/dev/null.
- Run the command and follow the shared block's reporting and resume-offer loop; the user resumes by saying "codex resume".
Quick Reference
| Use case |
Sandbox mode |
Key flags |
| Read-only review or analysis |
read-only |
--sandbox read-only 2>/dev/null |
| Apply local edits |
workspace-write |
--sandbox workspace-write --full-auto 2>/dev/null |
| Permit network or broad access |
danger-full-access |
--sandbox danger-full-access --full-auto 2>/dev/null |
| Resume recent session |
inherited from original |
echo "prompt" | codex exec --skip-git-repo-check resume --last 2>/dev/null (no flags) |
| Run from another directory |
match task needs |
-C <DIR> plus other flags 2>/dev/null |
Codex Specifics for the Shared Rules
The shared block leaves these caller-supplied:
- The peer CLI is Codex, powered by OpenAI models; its version check is
codex --version.
- High-impact flags that need permission first:
--full-auto, --sandbox danger-full-access, --skip-git-repo-check.
- Resume syntax for follow-ups and peer discussion of a disagreement:
echo "This is Claude (<your current model name>) following up. I disagree with [X] because [evidence]. What's your take?" | codex exec --skip-git-repo-check resume --last 2>/dev/null
Reference formatting
Before writing any output that mentions a file, standard, section, commit, or issue, read reference-formatting.md (RFC-0001) and apply it verbatim — link files, docs, skills, agents, and sections, and never leave a reference as bare text.
1---2name: ask-codex3description: Delegate a code analysis, refactoring, or automated-editing task to the OpenAI Codex CLI (codex exec / codex resume), then critically evaluate its output as a peer AI. Use when the user asks to run Codex, references OpenAI Codex, or wants a second model's take on the code.4---56# Ask Codex78Delegate a task to the OpenAI Codex CLI and report the result. Codex runs as a peer model: read [`peer-cli-delegation.md`](../shared-rules/references/peer-cli-delegation.md) and apply it throughout — it owns the delegation mechanics, the follow-up loop, the critical evaluation of Codex output, and the error-handling protocol, while this file supplies the CLI name, the exec command, the resume syntax, and the flag table. Do NOT use for: git/PR workflows (use the other autopilot skills) or tasks that do not involve the Codex CLI.910## Input1112Parse `$ARGUMENTS` for an optional task description and optional `--model` / `--effort` flags. Anything missing is collected in the next step.1314## Running a Task15161. If model or effort was not supplied in `$ARGUMENTS`, ask the user (via `AskUserQuestion`, **one prompt with two questions**) which model (`gpt-5.5`, `gpt-5.4`, `gpt-5.4-mini`, `gpt-5.3-codex-spark`, `gpt-5.3-codex`) and which reasoning effort (`xhigh`, `high`, `medium`, `low`).172. Select the sandbox mode; default to `--sandbox read-only` unless edits or network access are required.183. Assemble the command with the appropriate options:19 - `-m, --model <MODEL>`20 - `--config model_reasoning_effort="<xhigh|high|medium|low>"`21 - `--sandbox <read-only|workspace-write|danger-full-access>`22 - `--full-auto`23 - `-C, --cd <DIR>`24 - `--skip-git-repo-check`25 - `"your prompt here"` (final positional argument)264. Always pass `--skip-git-repo-check`.275. Resume: continue a prior session with `echo "your prompt here" | codex exec --skip-git-repo-check resume --last 2>/dev/null`. Insert any flags **between** `exec` and `resume`; per the shared block's inherit rule, pass no config flags on resume unless the user explicitly requests them.286. **IMPORTANT**: append `2>/dev/null` to every `codex exec` command to suppress thinking tokens (stderr). Only show stderr when the user explicitly asks to see thinking tokens or for debugging.297. **IMPORTANT (stdin)**: `codex exec` always reads stdin and concatenates it with the positional prompt — apply the shared block's stdin guard, e.g. `codex exec ... "prompt" </dev/null 2>/dev/null`.308. Run the command and follow the shared block's reporting and resume-offer loop; the user resumes by saying "codex resume".3132### Quick Reference3334| Use case | Sandbox mode | Key flags |35| ------------------------------ | ----------------------- | ---------------------------------------------------------------------------------------- |36| Read-only review or analysis | `read-only` | `--sandbox read-only 2>/dev/null` |37| Apply local edits | `workspace-write` | `--sandbox workspace-write --full-auto 2>/dev/null` |38| Permit network or broad access | `danger-full-access` | `--sandbox danger-full-access --full-auto 2>/dev/null` |39| Resume recent session | inherited from original | `echo "prompt" \| codex exec --skip-git-repo-check resume --last 2>/dev/null` (no flags) |40| Run from another directory | match task needs | `-C <DIR>` plus other flags `2>/dev/null` |4142## Codex Specifics for the Shared Rules4344The shared block leaves these caller-supplied:4546- The peer CLI is **Codex**, powered by OpenAI models; its version check is `codex --version`.47- High-impact flags that need permission first: `--full-auto`, `--sandbox danger-full-access`, `--skip-git-repo-check`.48- Resume syntax for follow-ups and peer discussion of a disagreement:4950```bash51echo "This is Claude (<your current model name>) following up. I disagree with [X] because [evidence]. What's your take?" | codex exec --skip-git-repo-check resume --last 2>/dev/null52```5354## Reference formatting5556Before writing any output that mentions a file, standard, section, commit, or issue, read [`reference-formatting.md`](../shared-rules/references/reference-formatting.md) (RFC-0001) and apply it verbatim — link files, docs, skills, agents, and sections, and never leave a reference as bare text.
awinogradov