AI glossary Operations
What is a skill catalog for AI agents?
← All glossary termsA governed registry of the skills AI agents can use, with names, versions, owners, required tools and dependencies.
Explained
A skill catalog is the registry where an organisation keeps track of the skills its AI agents can load: packaged instructions, scripts and resources that teach an agent how to do a specific task. For each skill it records what it does, who maintains it, which version is current, what tools and permissions it needs and how it combines with other skills. It prevents unmanaged prompts spreading across teams (shadow prompts), makes skills reusable and auditable, and fits with spec-driven change: the spec is updated first, then the skill package, then the deployments that use it.
What is an agent skill?
A skill is a reusable package that gives an agent specialised know-how. In the Agent Skills format introduced by Anthropic, a skill is a folder with a SKILL.md file of instructions plus optional scripts and resources. The agent sees a short description of each available skill and loads the full content only when the task calls for it, which keeps the context small.
As agents take on more work, the number of skills grows quickly, and so does the need to know which ones exist and which ones can be trusted.
What a skill catalog records
- Name, purpose and the description the agent uses to decide when to load it.
- Owner and maintainers.
- Version history and status: draft, approved, deprecated.
- Tools, connectors and permissions the skill requires.
- Dependencies on other skills or specs.
- Where it is deployed and which agents use it.
- Review and security checks, since skills can include executable code.
Why govern skills centrally
- Reuse: teams find and share skills instead of rewriting the same instructions.
- Security: skills from untrusted sources can run code or exfiltrate data, so they need review.
- Consistency: one approved version of how a task is done, not many copies drifting apart.
- Audit: you can trace which skill and version an agent used for a given action.
- Change control: updating a skill follows the same path as any other governed change.
Frequently asked questions
What are agent skills?
They are packaged instructions, scripts and resources that an AI agent can load when it needs to do a specific task, such as filling in a template or following a company procedure. They extend what a general-purpose agent can do without retraining the model.
What is the difference between a skill and a tool?
A tool is a function the agent can call, such as searching a database or sending an email. A skill is know-how: instructions on how to do a task, which may tell the agent which tools to use and in what order. Skills often depend on tools.
Why do you need a skill catalog?
Without one, skills and prompts multiply across teams with no owner, no versions and no review. A catalog makes it clear which skills exist, which are approved and who is responsible for each, which is essential for security and audit.
Are skills safe to install?
Only if they come from a trusted source and have been reviewed. Skills can contain executable code and instructions that change what an agent does, so a catalog should record who reviewed each skill and what permissions it needs.
How does a skill catalog relate to spec-driven development?
In a spec-driven approach, the specification of a process is the source of truth. When the process changes, the spec is updated first, then the skill that implements it, and then the agents that use it are redeployed. The catalog records each step.
Where it shows up at Thinkia
Further reading
Related terms
Sources
- Anthropic · Introducing Agent Skills
- Anthropic · Equipping agents for the real world with Agent Skills
- Agent Skills · Open specification for agent skills (SKILL.md)
Updated: 2026-10-21