wentorai/research-claw

codex-cli

Reference for using OpenAI Codex CLI to handle complex coding tasks. Use when Research-Claw needs to delegate multi-file projects, iterative debugging, or complex code generation to a specialized AI coding tool. Trigger words: codex, coding CLI, delegate coding, complex project.

First seen Apr 7, 2026

Installation

$ npx skills add wentorai/research-claw --skill codex-cli

Similar popular skills

Related neighbors and high-traction skills in the same topics — useful to compare before installing.

Also in this package

Other skills from wentorai/research-claw · top by installs.

npx skills add wentorai/research-claw

Browse all from wentorai/research-claw

More details

Agent compatibility

Declared targets from SKILL.md / docs. Unmarked agents are not listed — the skill may still install via the CLI.

Claude Code Not declared
Cursor Not declared
Codex Declared
GitHub Copilot Not declared
Windsurf Not declared
Gemini CLI Not declared
Cline Not declared
OpenCode Not declared

Repository health

Stars 851
License LICENSE
Default branch main
Open issues 0
Status Active

Skill metadata

Parsed from SKILL.md frontmatter.

Declared agents codex

Package contents

Files included with this skill beyond the listing page.

  • skill md SKILL.md 2,737 B
  • docs SUMMARY.md 296 B

History

  1. First seen on skills.sh
  2. First recorded snapshot · 10 installs

SKILL.md

Codex CLI (OpenAI)

Terminal-based AI coding agent by OpenAI. Reads, edits, and runs code locally in a sandboxed environment. Built in Rust.

Detect Installation

which codex  # should return a path
codex --version

Install

npm i -g @openai/codex

Requires an OpenAI API key or ChatGPT account (prompted on first run).

Non-Interactive Usage (exec)

For programmatic invocation from Research-Claw, use codex exec:

# Basic non-interactive execution
codex exec "Refactor the auth module to use JWT tokens"

# Pipe from stdin
echo "Fix the failing test in src/utils.test.ts" | codex exec -

# Full-auto mode (workspace writes + minimal approvals)
codex exec --full-auto "Add error handling to all API routes"

# Skip all approvals (dangerous — use only in sandboxed environments)
codex exec -a never -s workspace-write "Update dependencies and fix breaking changes"

# Save final response to file
codex exec -o result.md "Explain the architecture of this project"

Key Flags

Flag Description
-m, --model <model> Override model (e.g. gpt-5-codex)
-a, --ask-for-approval <mode> untrusted / on-request / never
-s, --sandbox <policy> read-only / workspace-write / danger-full-access
--full-auto Shortcut: workspace-write sandbox + on-request approvals
--add-dir <path> Grant write access to additional directories
-C, --cd <path> Set working directory for the agent
-i, --image <path> Attach image files (comma-separated)
-o, --output-last-message <path> Write final response to file
--json Print newline-delimited JSON events
--ephemeral Don't persist session to disk
--skip-git-repo-check Allow running outside a git repository
--oss Use local OSS model via Ollama

Research-Claw exec Patterns

# Delegate a complex coding task
exec("codex exec --full-auto -C /path/to/project 'Implement the feature described in TODO.md'")

# Read-only analysis
exec("codex exec -s read-only -o /tmp/analysis.md -C /path/to/project 'Analyze code quality and list issues'")

# With model selection
exec("codex exec -m gpt-5-codex --full-auto 'Fix all TypeScript type errors'")

Docs