Skills in OpenAI API

2026-02-11 · Source: Simon Willison's Weblog · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering · Depth: Intermediate, quick

Summary

OpenAI has integrated "Skills" directly into its API, allowing developers to utilize a shell tool for custom functionalities. These skills can be uploaded as zipped files or, more conveniently, sent inline within JSON requests as base64-encoded zip data. An example script demonstrates this by defining an "inline" skill named "wc" to count words in a file, using the `gpt-5.2` model. This capability was explored and documented through an automated research process using Claude Code for web use, specifically a tool called Showboat, which generated a detailed demo and report based on an initial prompt to fetch and analyze OpenAI's API documentation.

Key takeaway

For AI Engineers integrating custom logic into OpenAI applications, you should explore the new inline skills feature. This allows for dynamic, on-the-fly tool definition without pre-uploading, potentially streamlining development workflows and enabling more flexible agentic behaviors within your `gpt-5.2` model interactions. Consider using automated research tools to quickly prototype and understand new API capabilities.

Key insights

OpenAI API now supports inline, base64-encoded custom skills via a shell tool.

Principles

• Automated API exploration accelerates development.
• Inline skill definition simplifies deployment.

Method

Define skills as base64-encoded zip data within the OpenAI API JSON request for immediate use.

In practice

• Embed custom tools directly in API calls.
• Automate API research with AI agents.

Topics

• OpenAI API
• AI Skills Integration
• Shell Tool
• Base64 Data Embedding
• Claude Code Assistant

Code references

• simonw/research

Best for: AI Engineer, Machine Learning Engineer, Software Engineer

Related on AIssential

• dotnet / skills — The dotnet/skills repository centralizes pre-built .NET agent skills and custom agents adhering to the agentskills.io open standard.
• openai / openai-cookbook — The OpenAI Cookbook offers practical code examples and guides for using the OpenAI API.
• Closing the knowledge gap with agent skills — Agent skills effectively bridge LLM knowledge gaps for rapidly changing software development practices.
• Getting started with OpenAI Assistant APIs and Python — OpenAI Assistant APIs enable integration of custom GPTs with tailored instructions and knowledge into applications.
• xAI Releases Grok Skills and Updates Tool Calling Responses API — Grok Skills provides persistent, customizable AI expertise and integrated tool calling for automated workflows across platforms.
• OpenAI Launches Full-Scale Effort to Patch Open-Source Bugs as It Takes on Anthropic’s Mythos — OpenAI is significantly expanding its cybersecurity initiatives and tools to address AI hacking risks.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Simon Willison's Weblog.