Skill Building

Move skills between library workspaces, such as from a personal to a team library

Review a skill content and structure before choosing to keep, curate or remove it

Publish a managed skills library to GitHub and get a shareable install command

Update installed skills to latest versions after a dry run and breaking-change check

Turn local prompt folders into validated SKILL.md skills and sync them to agents

Design, review and test reusable SKILL.md files with precise trigger descriptions

Write evolving skill plans that capture steps, tools built and lessons learned

Audit installed skills, merge overlaps and generate a skill routing registry

List installed agent skills by plugin source and flag overlap with your own

Turn a proven workflow from a conversation into a reusable project skill

Check npm for a newer skill suite version and install it with a change summary

Turn a working conversation into a reusable agent skill you can save and share

Analyze skill execution traces and improve skills from real run data

Find and install skills from skills.sh that match a task you describe

Update an existing skill based on feedback about what should change

Plan which existing and new skills an agent workflow needs

Turn repeated tool sequences and code patterns into tested, reusable agent skills

Create and improve agent skills from sources and positive and negative examples

Turn your domain expertise into a SKILL.md and submit it for maintainer review

Walk through contributing a new skill to a GitHub repo as a pull request

Turn cognitive science papers into structured, reusable methods as skills

Convert a GitHub repository or local codebase into a reusable agent skill

Share a research case from your session to the community on GitHub Discussions

Share anonymized statistics on which skills get used most with the community

Verify a skill for parameter accuracy, citations and methodology through expert review

Create a custom agent persona, restricted specialist or project-local skill

Distill concept and method books into traceable task skills with human review gates

Create, test and refine agent skills with validation scripts and trigger tuning

Find which discipline skill fits a scientific task across 23 benchmark domains

Turn a verified research session into a validated skill candidate with rollback

Track which agent skills work and propose improvements to the skill library