The infrastructure layer for AI tooling
A browsable catalog and machine-readable API that answers the question: "Which skill should I use?"
What It Does
Skill Directory solves a simple but critical problem: when an AI agent needs help, how does it find the right tool? Myles built a dual-purpose infrastructure layer -- a clean landing page for humans to browse available Agent Skills, paired with a JSON API that agents can query directly.
The landing page lists each skill with its name, description, use cases, and links to repo and demo. For agents, the GET /api/skills endpoint returns the same data in machine-readable format. Humans get discovery; agents get direct access.
Why Skill Directory Matters
Week 2's theme was "tools for AI tools." Most builders focused on building individual skills -- useful capabilities agents could invoke. Myles stepped back and asked: what's the infrastructure those skills sit on?
The Skill Directory is that infrastructure. It's the layer that makes "which skill do I use?" answerable. Build once, and any agent (or human) can discover what's available across the ecosystem.
Key insight: The directory includes three actual skills from the community -- Myles' own sprint-helper, Kyle Sebestyen's SQL Server profiler, and Jake Strait's crypto market skill -- making it immediately useful, not just theoretical.
Community Shout-outs
Myles specifically highlighted complementary work from other competitors:
- Jake Strait's Crypto Market Agent Skill -- live crypto prices and market data in JSON, with a clean demo at crypto-agent-skill.vercel.app
- Kyle Sebestyen's SQL Server DB Profiler -- an agent-invokable tool that surfaces database metrics (DTUs, top queries, tuning recommendations) directly in the IDE using PowerShell and SQL scripts mapped to natural-language input
This cross-pollination is exactly what the sprint is designed for -- competitors building on and amplifying each other's work.
Links
- Live Demo: mf-spring-into-ai-week2.vercel.app
- API Endpoint: mf-spring-into-ai-week2.vercel.app/api/skills
- GitHub: github.com/myles-mf/mf-spring-into-ai
- Original Post: @merlinforge on Substack