By shareai-lab
Extend AI agents with specialized skills to design and build custom agents for various domains, transform technical content into platform-specific social posts, evaluate skill designs against best practices, and collaborate on disciplined coding for features, bugs, and refactoring.
npx claudepluginhub shareai-lab/shareai-skillsDesign and build AI agents for any domain. Use when users: (1) ask to "create an agent", "build an assistant", or "design an AI system" (2) want to understand agent architecture, agentic patterns, or autonomous AI (3) need help with capabilities, subagents, planning, or skill mechanisms (4) ask about Claude Code, Cursor, or similar agent internals (5) want to build agents for business, research, creative, or operational tasks Keywords: agent, assistant, autonomous, workflow, tool use, multi-step, orchestration
Create platform-native content that resonates with each community's culture. Use when adapting technical content for WeChat, Hacker News, Reddit, Medium, Twitter/X, Dev.to, or LinkedIn. Transforms generic writing into content that feels written BY that community, not AT them.
Evaluate Agent Skill design quality against official specifications and best practices. Use when reviewing, auditing, or improving SKILL.md files and skill packages. Provides multi-dimensional scoring and actionable improvement suggestions.
Transform an AI agent into a tasteful, disciplined development partner. Not just a code generator, but a collaborator with professional standards, transparent decision-making, and craftsmanship. Use for any development task: building features, fixing bugs, designing systems, refactoring. The human provides vision and decisions. The agent provides execution with taste and discipline.
Knowledge packages that extend AI agent capabilities.
Works with Kode CLI, Claude Code, Cursor, and any agent supporting the Agent Skills Spec.
kode plugins install https://github.com/shareAI-lab/shareAI-skills
claude plugins install https://github.com/shareAI-lab/shareAI-skills
Copy the skills/ directory to your Cursor skills folder.
Load SKILL.md files on-demand when the agent needs domain expertise.
| Skill | Description |
|---|---|
| skill-judge | Evaluate Agent Skill quality across 8 dimensions (120-point system) |
| media-writer | Adapt content for WeChat, HN, Reddit, Medium, Twitter, Dev.to, LinkedIn |
| agent-builder | Design and build AI agents for any domain |
| vibe-coding | Vibe-driven development with minimal specs |
Creating a truly effective skill is an art. We've analyzed 17 official Anthropic skills and distilled the core principles:
Core Formula:
Good Skill = Expert-only Knowledge - What Claude Already Knows
Read the full guide: How to Create Great Agent Skills
Use the skill-judge skill to evaluate your skill's quality with structured scoring across 8 dimensions:
| Dimension | Points | Focus |
|---|---|---|
| Knowledge Delta | 20 | Expert-only knowledge vs. what Claude already knows |
| Mindset + Procedures | 15 | Thinking patterns + domain-specific workflows |
| Anti-Pattern Quality | 15 | Specific NEVER lists with non-obvious reasons |
| Specification (esp. Description) | 15 | Description with WHAT, WHEN, and KEYWORDS |
| Progressive Disclosure | 15 | Content layering, loading triggers |
| Freedom Calibration | 15 | Specificity matched to task fragility |
| Pattern Recognition | 10 | Follows established skill patterns |
| Practical Usability | 15 | Decision trees, working examples, edge cases |
Skills are modular knowledge packages that give AI agents domain expertise on-demand. They follow the Agent Skills Spec.
skill-name/
├── SKILL.md # Core instructions (required)
├── references/ # Detailed documentation (optional)
├── scripts/ # Executable code (optional)
└── assets/ # Templates and resources (optional)
Skills are knowledge, not code.
A skill doesn't tell the agent what to do step-by-step. It gives the agent the knowledge to figure out what to do. The model is smart - your job is to inform it, not constrain it.
We welcome contributions! To add a new skill:
skills/SKILL.md with YAML frontmatter (name, description)| Repository | Purpose |
|---|---|
| Kode | Full-featured open source agent CLI |
| learn-claude-code | Learn how to build AI agents from scratch |
| Agent Skills Spec | Official specification |
Apache-2.0
Open collection of AI agent skills — reusable, framework-agnostic SKILL.md packages
Share bugs, ideas, or general feedback.
Ultimate Claude Code skill creator. Design, scaffold, build, review, evolve, and publish production-grade AI agent skills following the Agent Skills open standard and 3-layer architecture.
Professional skill and subagent creation with dual-mode workflow: 12-step fast mode and 15-step full mode with behavioral pressure testing and TDD integration.
Helps users discover and install agent skills when they ask questions like "how do I do X", "find a skill for X", "is there a skill that can...", or express interest in extending capabilities. This is the primary skill for finding, searching, installing, and managing AI agent skills. Use /learn to search 100,000+ skills from agentskill.sh, install them mid-session, track versions, and rate skills after use. This skill should be used when the user is looking for functionality that might exist as an installable skill.
Create and validate production-grade agent skills with 100-point marketplace grading
Ultra-compressed communication mode. Cuts ~75% of tokens while keeping full technical accuracy by speaking like a caveman.