Skills
Extend what your AI can do with modular skills — like apps on a phone, each one adds a new capability.
Skills are modular capabilities that extend what your AI assistant and agents can do. Think of them like apps on a phone — your phone is useful on its own, but apps make it powerful for specific tasks. Skills work the same way for your AI.
What skills do
A skill gives the AI the ability to do something specific. Without skills, your AI can chat, answer questions, and reason through problems. With the right skills, it can:
- Search and analyze documents in your knowledge base
- Generate reports from structured data
- Draft emails and messages in your preferred style
- Process files in specific formats
- Connect to external tools and services
Each skill is self-contained. You can mix and match them to build exactly the AI you need.
Skill scopes
Skills are organized into three scopes, depending on who can use them:
- Common — Built-in skills available to every assistant and agent. These cover the basics that everyone needs.
- Organization — Skills created by your team and shared across your workspace. Anyone in your organization can use them.
- Agent — Custom skills built for a specific agent. These are tailored to that agent's role and aren't shared with others.
When you create a new agent, it automatically has access to all common skills and your organization's shared skills. You can then add agent-specific skills to give it unique capabilities. For more on how agents and skills work together, see Agents and Skills.
How skills are structured
Every skill has three core pieces:
- Name — A clear, descriptive label like "Document Summarizer" or "Email Drafter"
- Description — A short explanation of what the skill does, so the AI knows when to use it
- Instructions — Detailed guidance that tells the AI exactly how to perform the skill
The description is especially important. The AI reads it to decide which skill to use for a given task, so a well-written description means your skills get used at the right time.