ask-codex

Ask Codex

Consult OpenAI Codex (GPT-5) as a second opinion for investigation, debugging, or review tasks. Codex runs in read-only mode with full project access — it analyzes, Claude implements.

Activation Triggers

• "ask codex", "check with codex", "codex review"
• "what does codex think", "get codex opinion"
• "consult codex", "run codex on this"

Workflow

Step 1: Check Availability

Run which codex to verify the CLI is installed. If not found, inform the user and stop.

Step 2: Build Context

Gather context from the current conversation:

1. What's the problem/question — summarize in 2-3 sentences
2. What we know — relevant files, error messages, behavior observed
3. What we tried — approaches attempted and why they failed (if applicable)
4. Specific question — what exactly codex should analyze or answer

Step 3: Construct Prompt

Build a focused prompt. Do NOT dump entire files — codex has full project access and can read them itself. Provide file paths and line references so codex knows where to look.

Template:

# [Investigation/Debug/Review] Request

## Problem
[2-3 sentence description]

## Context
- Files: [path/to/file.go:lineNumber, ...]
- Observed: [what's happening]
- Expected: [what should happen]

## What We Tried
[List approaches and outcomes, or "First consultation" if fresh question]

## Question
[Specific, focused question for codex to answer]

Provide:
1. Root cause analysis (if debugging)
2. Concrete recommendation with file:line references
3. Why previous approaches failed (if applicable)

Keep response focused and actionable.

Step 4: Execute Codex

Run codex in background (it takes 2-5 minutes for complex analysis):

codex exec -m gpt-5 \
  --sandbox read-only \
  -c model_reasoning_effort="high" \
  -c stream_idle_timeout_ms=600000 \
  -c project_doc="./CLAUDE.md" \
  "prompt here"

Execution rules:

• Always use run_in_background: true in Bash tool
• Monitor with BashOutput every 15-20 seconds
• Be patient during reasoning phase (1-3 minutes of silence is normal)
• Total timeout: 10 minutes for standard, 15 for complex

Flags:

• --sandbox read-only — codex can read all project files but cannot modify anything
• -m gpt-5 — codex model (adjust to latest available)
• model_reasoning_effort="high" — maximum reasoning depth
• project_doc — passes CLAUDE.md as project context

Step 5: Present Results

1. Extract codex's analysis — skip session info, token counts, prompt echo
2. Present findings clearly formatted
3. Add assessment — agree, disagree, or note caveats
4. Propose next steps — ask if user wants to implement the suggestion

Output format:

**Codex Analysis:**

[Codex's response — cleaned up and formatted]

---

**Assessment:** [2-3 sentence evaluation of codex's findings]

**Next steps:** [What to do with this information]

Important Rules

• Read-only always — codex analyzes, Claude implements. Never let codex edit files.
• Don't duplicate files — codex has full project access. Provide paths, not content.
• Focused prompts — specific questions get better answers than broad "review everything".
• Background execution — always run in background to avoid timeout issues.
• One question at a time — if multiple concerns, run separate codex queries.
• Critical thinking — codex can be wrong. Evaluate its suggestions before implementing.