Using skills to accelerate OSS maintenance | OpenAI Developers
Mar 9, 2026

Summary
This post details how OpenAI uses Codex and repo-local skills (`AGENTS.md`, `.agents/skills/`, GitHub Actions) to automate and accelerate open-source maintenance for the OpenAI Agents SDK repositories. It introduces specific skills for code verification, documentation synchronization, example execution, release review, implementation strategy, knowledge retrieval from OpenAI docs, and PR draft summarization. The post emphasizes the pattern of narrow skill contracts, clear triggers, and concrete outputs, and how `AGENTS.md` enforces mandatory skill usage for critical workflows like code change verification and changeset validation in the JavaScript SDK.
More from the Agentic Execution Environment trail
- Bringing my LED display to life with GPT-Live-1 and Codex | OpenAI Developers
- Rethinking skills and prompts for GPT-6 Astra | OpenAI Developers
- Meet Rosalind Workbench: Empowering every scientist to be their own research team | OpenAI Developers
- Automating repetitive work at OpenAI with Codex
- Codex as a platform: build on the open agent harness | OpenAI Developers