From local-codex
Launch and manage OpenAI Codex CLI as a non-interactive coding sub-agent for code generation, refactoring, and review using ChatGPT Pro OAuth.
How this skill is triggered — by the user, by Claude, or both
Slash command
/local-codex:local-codexThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Delegate coding tasks to the local OpenAI Codex CLI agent using your ChatGPT Pro subscription (OAuth, no API charges).
Delegate coding tasks to the local OpenAI Codex CLI agent using your ChatGPT Pro subscription (OAuth, no API charges).
codex exec for non-interactive tasksCRITICAL: This skill uses OAuth authentication from ~/.codex/auth.json (ChatGPT Pro flat-rate subscription). Do NOT set OPENAI_API_KEY or pass API keys — that would switch to pay-per-use billing.
codex login in terminal (browser OAuth flow)The skill auto-detects Codex CLI in this order:
/Applications/Codex.app/Contents/Resources/codex (desktop app)/usr/local/bin/codex (npm global)/opt/homebrew/bin/codex (Homebrew)~/.npm-global/bin/codexwhich codex fallbackpython3 scripts/codex_wrapper.py exec "<prompt>" [<workdir>] [<model>] [<sandbox>] [<timeout>]
Example:
python3 scripts/codex_wrapper.py exec \
"Write a Python function to calculate fibonacci" \
/tmp \
gpt-5.5 \
workspace-write \
300
python3 scripts/codex_wrapper.py review [<workdir>] [<model>] [uncommitted:true] [<timeout>]
Example:
python3 scripts/codex_wrapper.py review \
/path/to/repo \
gpt-5.5 \
true \
300
python3 scripts/codex_wrapper.py status
The wrapper returns JSON with:
success: boolexit_code: intelapsed_seconds: floatstdout: raw outputstderr: error stream (truncated)parsed_jsonl: parsed JSONL events (if --json)final_message: extracted assistant text (if available)| Parameter | Default | Options |
|---|---|---|
| model | gpt-5.5 | gpt-5.5, gpt-5.5-pro, o4-mini, etc. |
| sandbox | workspace-write | read-only, workspace-write, danger-full-access |
| timeout | 300 | seconds (increase for large tasks) |
| json_output | true | always true (wrapper parses JSONL) |
sandbox=read-only for analysis/review tasks (no file writes)sandbox=workspace-write for code generation (writes to working dir)sandbox=danger-full-access only when explicitly needed (full system access)--skip-git-repo-check when running outside git repos--ephemeral for one-off tasks (no session persistence)For multi-step tasks, Codex supports session resume:
# First step
codex exec --ephemeral "Step 1..."
# Later
codex exec resume --last "Step 2..."
The wrapper currently runs single-shot exec. For multi-step workflows, use raw codex exec commands.
codex doctor is buggy in current version (alpha), avoid usingnpx claudepluginhub p/daymade-local-codex-local-codexUse OpenAI Codex from inside Claude Code for code reviews and delegated background tasks.
Launches Codex CLI as an isolated subagent for bounded coding, review, or verification tasks in a separate session with clear success criteria.
Delegates coding tasks to OpenAI Codex CLI (GPT-5.3-codex) for second opinions, sandboxed execution, and automated code editing. Supports session resume/fork and structured JSON output.