codex

Codex Skill Guide

Running a Task

1. Ask the user (via AskUserQuestion) which model to run (gpt-5.4, gpt-5.3-codex-spark, gpt-5.3-codex, or gpt-5.2) AND which reasoning effort to use (xhigh, high, medium, or low) in a single prompt with two questions. Default to gpt-5.4 with high reasoning if not specified.
2. Select the sandbox mode required for the task; default to --sandbox read-only unless edits or network access are necessary.
3. 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>
   • --add-dir <DIR> (additional directories that should be writable alongside the primary workspace)
   • -i, --image <FILE> (attach image(s) to the prompt)
   • --ephemeral (run without persisting session files to disk)
   • --skip-git-repo-check
   • "your prompt here" (as final positional argument)
4. Always use --skip-git-repo-check.
5. When continuing a previous session, use codex exec --skip-git-repo-check resume --last via stdin. When resuming don't use any configuration flags unless explicitly requested by the user e.g. if they specify the model or the reasoning effort when requesting to resume a session. Resume syntax: printf '%s' "your prompt here" | codex exec --skip-git-repo-check resume --last 2>/dev/null. All flags have to be inserted between exec and resume.
6. IMPORTANT: By default, append 2>/dev/null to all codex exec commands to suppress thinking tokens (stderr). Only show stderr if the user explicitly requests to see thinking tokens or if debugging is needed.
7. For prompts containing special characters, use printf for safe piping:

   printf '%s' "Your prompt with 'quotes' and $pecial chars" | codex exec -m <MODEL> -c model_reasoning_effort="high" --sandbox read-only --skip-git-repo-check 2>/dev/null
   

8. Always use run_in_background: true on the Bash tool when running codex exec. Codex tasks routinely take 5-15+ minutes. Do NOT use shell & or inline waiting. The system will notify you automatically when the command completes. Continue with independent work while waiting, but do NOT skip the Codex results — they are load-bearing.
9. Run the command, capture stdout/stderr (filtered as appropriate), and summarize the outcome for the user.
10. After Codex completes, inform the user: "You can resume this Codex session at any time by saying 'codex resume' or asking me to continue with additional analysis or changes."

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
Code review (built-in)	read-only	codex exec review 2>/dev/null
Multi-directory writes	workspace-write	--sandbox workspace-write --add-dir <DIR> --full-auto 2>/dev/null
With image context	Match task needs	-i screenshot.png "Analyze this UI" 2>/dev/null
Ephemeral (no session saved)	Match task needs	--ephemeral 2>/dev/null
Resume recent session	Inherited from original	printf '%s' "prompt" | codex exec --skip-git-repo-check resume --last 2>/dev/null
Run from another directory	Match task needs	-C <DIR> plus other flags 2>/dev/null

Following Up

• After every codex command, immediately use AskUserQuestion to confirm next steps, collect clarifications, or decide whether to resume with codex exec resume --last.
• When resuming, pipe the new prompt via stdin: printf '%s' "new prompt" | codex exec --skip-git-repo-check resume --last 2>/dev/null. The resumed session automatically uses the same model, reasoning effort, and sandbox mode from the original session.
• Restate the chosen model, reasoning effort, and sandbox mode when proposing follow-up actions.

Critical Evaluation of Codex Output

Codex is powered by OpenAI models with their own knowledge cutoffs and limitations. Treat Codex as a colleague, not an authority.

Guidelines

• Trust your own knowledge when confident. If Codex claims something you know is incorrect, push back directly.
• Research disagreements using WebSearch or documentation before accepting Codex's claims. Share findings with Codex via resume if needed.
• Remember knowledge cutoffs - Codex may not know about recent releases, APIs, or changes that occurred after its training data.
• Don't defer blindly - Codex can be wrong. Evaluate its suggestions critically, especially regarding:
  • Model names and capabilities
  • Recent library versions or API changes
  • Best practices that may have evolved

When Codex is Wrong

1. State your disagreement clearly to the user
2. Provide evidence (your own knowledge, web search, docs)
3. Optionally resume the Codex session to discuss the disagreement. Identify yourself as Claude so Codex knows it's a peer AI discussion. Use your actual model name (e.g., the model you are currently running as) instead of a hardcoded name:

   printf '%s' "This is Claude (<your current model name>) following up. I disagree with [X] because [evidence]. What's your take on this?" | codex exec --skip-git-repo-check resume --last 2>/dev/null
   

4. Frame disagreements as discussions, not corrections - either AI could be wrong
5. Let the user decide how to proceed if there's genuine ambiguity

Error Handling

• Stop and report failures whenever codex --version or a codex exec command exits non-zero; request direction before retrying.
• Before you use high-impact flags (--full-auto, --sandbox danger-full-access, --skip-git-repo-check) ask the user for permission using AskUserQuestion unless it was already given.
• When output includes warnings or partial results, summarize them and ask how to adjust using AskUserQuestion.