Agent Skills

Researches GitHub issues and pull requests from a local archive before escalating to live repository queries.

Builds, tests, and refines Agent Skills through an evaluation loop rather than treating SKILL.md as static prose.

Searches the open Agent Skills ecosystem before an agent reinvents a workflow that already exists.

Guide for creating Model Context Protocol servers that enable LLMs to interact with external services through well-designed tools

Create stateful coordination systems on Cloudflare with Durable Objects for chat rooms, multiplayer games, booking systems, and real-time collaboration

Build stateful AI agents on Cloudflare Workers using the Agents SDK with durable execution, state management, and real-time capabilities

A guarded workflow for diagnosing failed GitHub Actions checks, extracting useful logs, and implementing a fix only after approval.