Platforms
Claude Code Skills
This page uses Anthropic documentation as the baseline for Claude Code skills, with emphasis on `.claude/skills`, installation flow, and common engineering tasks.
What you'll find here
Anthropic describes Claude Code skills as reusable capabilities for recurring tasks. A skill can include more than instructions alone: scripts, example files, and operating notes can all be packaged together so Claude Code can reuse a stable workflow.
The official docs also make the install paths clear. Project-specific skills live in `.claude/skills/`, while user-wide defaults live in `~/.claude/skills/`. For teams, the project path is usually the better fit for release, review, and test flows that evolve with the repository. The user path is more useful for personal workflows you want across many projects.
That is why this page works best for people who already know they want Claude Code. Start here to narrow the field to Agent Skills that genuinely fit Claude Code conventions, then open the skill detail pages to verify maintenance signals, dependencies, and setup assumptions.
If you are still sorting out the boundary between skills, MCP, and other extension patterns, read the Anthropic docs and the install guide first. Once the model is clear, it becomes much easier to choose the right Claude Code skills instead of collecting a pile of mismatched experiments.
Official references
Related skills
Start with the skills that are closest to this topic and easiest to evaluate.


