Skills Protocol

A minimal protocol for connecting LLM agents to Skills (instructions + code) and Blobs (large data) via a Skills Runtime with sandboxed code execution.

Overview

The Skills Protocol is a JSON-RPC 2.0 over HTTP specification that provides a standardized way for LLM agents to discover, inspect, and execute skills. Instead of exposing each skill as a native LLM tool, the protocol provides a small set of meta-tools that let agents work with an extensible skill ecosystem.

Key features

Getting started

The Skills Protocol exposes 8 core tools to LLM agents:

1. list_skills — enumerate available skills
2. describe_skill — retrieve skill manifest and documentation
3. read_skill_file — read any file in a skill's directory
4. execute_skill — run a skill's entrypoint
5. run_code — execute LLM-written code with mounted skills
6. create_blob — store large content
7. read_blob — retrieve blob previews
8. load_skills_protocol_guide — bootstrap instruction tool