zjunlp/SkillNet

Create, Evaluate, and Connect AI Skills

What it solves

SkillNet provides an infrastructure layer for the reuse of AI agent skills, preventing developers from having to rebuild the same capabilities from scratch. It treats agent skills as software assets that can be searched, installed, and composed.

How it works

SkillNet operates as a library and ecosystem that allows users to:

  • Discover and Install: Search a public skill library via keywords or semantic intent (vector search) and download skill folders directly from GitHub into local workspaces.
  • Create: Generate structured skill packages from existing repositories, documents, prompts, or execution traces.
  • Evaluate: Score skills based on safety, completeness, executability, maintainability, and cost awareness.
  • Analyze and Compose: Infer relationships (such as dependencies or handoffs) between local skills to create scenario graphs.
  • Orchestrate: Select specific skills for a given scene (e.g., "sciatlas") and generate a prompt for a downstream execution agent.

Who it’s for

AI agent developers and researchers who want to integrate pre-built capabilities into their agents without manual implementation, as well as those building agent runtimes like Claude Code or OpenClaw.

Highlights

  • Public Skill Library: Access to over 500K GitHub skills with a dedicated visual explorer.
  • Portable Packages: Ability to turn documents and prompts into installable skill packages.
  • Comprehensive Tooling: Includes a Python SDK, CLI, and REST API for seamless integration.
  • Agent Runtime Integration: Built-in support for MCP (Model Context Protocol), Claude Code, and Codex.
  • Scenario-based Orchestration: Ability to map tasks to specific skill collections and generate execution prompts.

Related

  • Project
    chujianyun/skills

    A curated library of reusable Agent Skills and workflows for AI agents, covering documentation retrieval, prompt optimization, career coaching, and system operations.

  • Project
    managedcode/dotnet-skills

    A catalog and management tool that provides modern .NET implementation guidance to AI agents, preventing them from suggesting outdated patterns.

  • Project
    dotnet/skills

    A curated collection of core skills and custom agents for coding assistants to perform specialized .NET development, migration, and diagnostic tasks.

  • Project
    HKUDS/OpenSpace

    A skill management layer for AI agents that enables them to retrieve, evaluate, and evolve their capabilities based on real-world task outcomes.

  • Project
    yaojingang/yao-meta-skill

    A governance and engineering system that transforms rough workflows and prompts into reusable, cross-platform AI agent skill packages with built-in evaluation and release gates.