awesome-codex-subagents
A collection of 130+ specialized Codex subagents covering a wide range of development use cases.
What is awesome-codex-subagents?
A collection of 130+ specialized Codex subagents covering a wide range of development use cases.
Overview of awesome-codex-subagents
What it is
This repository is a curated collection of over 176 Codex subagents organized into 13 categories. Each subagent is a specialized AI assistant defined in a Codex-native .toml format, designed for specific development tasks. The collection is written specifically for Codex and aligned with the official documentation.
Key features
- Over 176 subagents across 13 categories, including core development, language specialists, and infrastructure.
- Each subagent uses a
.tomlformat with fields for name, description, model, sandbox mode, and instructions. - Smart model routing: subagents automatically use models like
gpt-6.1-solfor complex tasks orgpt-6-lunafor bounded search and drafting. - Sandbox mode philosophy: subagents are either read-only (reviewers, auditors) or workspace-write (developers, engineers) to control filesystem access.
- Subagents can be installed globally (
~/.codex/agents/) or per-project (.codex/agents/), with project-specific agents taking precedence.
Use cases
- Everyday coding tasks: API design, backend development, frontend UI/UX, full-stack feature development.
- Language-specific development: experts for Python, JavaScript, TypeScript, Go, Rust, Java, C++, and many more.
- Infrastructure and DevOps: cloud architecture, Docker, Kubernetes, CI/CD, security auditing, incident response.
- Quality and security: code review, security auditing, diagnostics, and professional analysis.
Getting started
bashmkdir -p ~/.codex/agentscp categories/01-core-development/backend-developer.toml ~/.codex/agents/``````bashmkdir -p .codex/agentscp categories/04-quality-security/reviewer.toml .codex/agents/See the repository README for full installation instructions.
License and pricing
License: MIT. Pricing: open source.
Based on the project's public README and repository metadata.