Lesson 881 of 2116
What A Skill Is In OpenClaw: Anatomy And Discovery
OpenClaw skills are pluggable capabilities — manifest plus procedure plus examples — that a soul discovers and invokes when the job calls for them. Understanding the anatomy is the first step to building or auditing one. Skills are how an OpenClaw agent grows hands OpenClaw is an open-source agentic framework that runs on your own machine.
Lesson map
What this lesson covers
Learning path
The main moves in order
- 1Skills are how an OpenClaw agent grows hands
- 2skill
- 3manifest
- 4procedure
Concept cluster
Terms to connect while reading
Section 1
Skills are how an OpenClaw agent grows hands
OpenClaw is an open-source agentic framework that runs on your own machine. Out of the box, an OpenClaw soul — the term OpenClaw uses for an active agent instance — can read files, run commands, and talk to its model. Anything beyond that is a skill: a self-contained bundle of instructions and tools that the soul loads on demand when a request matches.
If you have used Claude Code skills or MCP servers, the shape will feel familiar. The difference is that OpenClaw skills run inside your local soul, are described in plain text the model reads, and discover themselves through trigger descriptions rather than through a custom UI.
The three parts of a skill
- 1Manifest: a small frontmatter block with name, description, and any permissions or trigger keywords. The description is what the soul matches against the user's request.
- 2Procedure: the natural-language instructions the soul follows once the skill is loaded — what steps to run, what tools to call, what output to produce.
- 3Examples and assets: optional templates, sample inputs, helper scripts, or reference data the procedure can point at when it needs concrete shape.
How a soul discovers a skill
When you start a session, OpenClaw scans the configured skill directories and reads each manifest. Only the manifests load into the system prompt at first — not the full procedures. The soul sees a catalog of available skills with their trigger descriptions. When the user's request rings a match, the soul loads the matching skill's procedure into context and follows it.
Invocation: explicit and implicit
Compare the options
| Mode | How it fires | When to use |
|---|---|---|
| Implicit (lazy) | Soul matches request to skill description and pulls in the procedure | Common workflows where users phrase requests naturally |
| Explicit (named) | User types `/skill-name` or names it directly | Skills with side effects, multiple matches, or ambiguous triggers |
| Forced (config) | Configuration pins certain skills to always load | Workspace-wide rules you never want skipped |
Where skills live
- User-scoped — in your home directory, available to every soul you spawn
- Workspace-scoped — checked into a project, version-controlled with your team
- Plugin-scoped — shipped by a third-party plugin or registry, namespaced by source
- Built-in — bundled with the OpenClaw distribution itself
Key terms in this lesson
The big idea: an OpenClaw skill is a manifest plus a procedure plus optional examples, discovered lazily by a soul that scans descriptions and loads only what matches. The description is the resume; the procedure is the interview.
End-of-lesson quiz
Check what stuck
15 questions · Score saves to your progress.
Tutor
Curious about “What A Skill Is In OpenClaw: Anatomy And Discovery”?
Ask anything about this lesson. I’ll answer using just what you’re reading — short, friendly, grounded.
Progress saved locally in this browser. Sign in to sync across devices.
Related lessons
Keep going
Creators · 9 min
OpenClaw: Souls, Heartbeats, And Skills
OpenClaw is an open-source agentic framework built around three primitives — souls (persistent personas with memory), heartbeats (autonomous loops), and skills (pluggable capabilities). Knowing those three tells you when OpenClaw is the right fit.
Creators · 9 min
Skills: Bundled Procedural Knowledge
Skills are reusable bundles of instructions plus optional scripts and assets. They're how Claude Code learns a procedure once and reapplies it everywhere.
Creators · 9 min
What Perplexity Is: Search-Augmented LLM, Not A Chatbot
Perplexity is built around the idea that every answer should cite its sources. Treating it like ChatGPT misses the point — and the reliability gap that comes with it.