A skill is a portable file you drop into Claude Code, Cursor, Windsurf, Copilot, or any AI coding assistant — and immediately gain a repeatable capability. Every skill in the Forge includes two things you won't find in a generic tutorial: why it works (the rationale — what makes this reliable and when to reach for it) and how it works (concrete steps and prompt patterns you can adapt to your own project).
Browse by category, download in the format your tool expects, and check the community compatibility notes before you commit. Built something worth sharing? Contribute it to the Forge.
Long-Running Task Checkpointer
Wrap an agent loop so every step snapshots state to disk, letting a crashed agent resume from the last good state instead of restarting from scratch.
Tool-Call Budget Guard
Cap an agent's tool calls with per-tool budgets and graceful degradation when exhausted — so runaway agents don't blow through API quotas.
Self-Correcting Schema Agent
An agent that retries JSON generation with schema violation feedback until the output validates — or a budget is exhausted.
Browser Testing Agent Loop
Agentic loop that exercises a web app like a new user, files bug reports with screenshots + reproduction steps.