Skip to content
Book a call
AI for go-to-market

Claude SkillsAgent Skills

Federico DonatoneBy · Updated October 15, 2026
Definition

Claude Skills, also called Agent Skills, are folders that teach Claude how to do a specific task. Each one holds a SKILL.md file with instructions, plus optional scripts and resources. Claude loads a skill only when a task calls for it, in Claude Code, the Claude apps and the API, so the same know-how works everywhere.

How Claude Skills work

A skill is a folder. At its center is SKILL.md, a plain-text file that opens with a short name and a description of when to use the skill, followed by the instructions themselves. The folder can also hold:

  • Scripts: small programs for steps that must be exact, such as counting DNS lookups or finding duplicate rows.
  • Resources: templates, worked examples and reference data the instructions point to.

Claude first sees only each skill's name and description. When a task matches, it reads the full instructions, and it opens the extra files only if the instructions call for them. That keeps unused skills out of the way, so you can install many without crowding the model's context. In Claude Code you can also call a skill directly with a slash command.

Claude Skills vs CLAUDE.md vs MCP

SkillCLAUDE.md rules fileMCP server
What it holdsKnow-how for one taskProject rules and contextA connection to a tool or data source
When it is usedWhen a task matches itLoaded at the start of every sessionWhen Claude calls the tool
ExampleGrade a cold emailOur voice and what never to doRead deals from the CRM

They work together. A skill can tell Claude which connected tool to use and in what order, while the rules file sets the limits every skill has to respect. A useful test: if the instruction applies to every task, it belongs in the rules file. If it applies to one job, make it a skill.

How to create a Claude skill

  1. Pick one task you repeat, with a result you can check, such as grading a cold email before it goes out.
  2. Do the task once with Claude, correcting it until the output is right.
  3. Ask Claude to save the working steps as a skill. It writes the SKILL.md file for you.
  4. Make the description specific: say what the skill does and the requests that should trigger it, in the words you would use.
  5. Add a script for any step that must be exact, and one example of excellent output.
  6. Test it on a new case, and add a rule each time it gets something wrong.
From our workThe description is the part people get wrong. If it is vague, Claude will not load the skill when it should. Write it as a trigger, not a title.

Free Claude skills for GTM teams

Our free GTM diagnostic skills for Claude are a working example. They are open source and install as a Claude Code plugin from our free tools page. Each skill checks one part of an outbound program and returns a score with the evidence behind it:

  • Cold email domain check: MX, SPF, DKIM, DMARC and blocklist status for a sending domain.
  • Cold email grader: checks one email and quotes the sentences that lose points.
  • Lead list QA: flags duplicates, invalid and role emails, and domains with no mail server.
  • Meeting show rate benchmark: compares your booked meetings with the 1,380 B2B meetings booked by cold calling in the Growth Cab show rate study.

To see how skills fit into a full agent with rules, tools and a schedule, read our guide to building an AI agent with Claude. For sales use cases, see Claude for sales.

Full guideBuild an AI Agent With Claude: Skip the Agent Apps

Claude Skills FAQ

Where can I use Claude Skills?
Skills work in Claude Code, in the Claude apps and through the Claude API. The same skill folder can move between them, so a skill you build for yourself in Claude Code can later serve your whole team. In Claude Code you can also call a skill by name with a slash command.
What is the difference between a Claude skill and a prompt?
A prompt is typed once and lives only in that conversation. A skill is saved as a file, loads itself when a matching task comes up and can include scripts and examples. Turn a prompt into a skill when you catch yourself pasting it a second time.
Do I need to code to write a Claude skill?
No. SKILL.md is plain English, and Claude can write the file for you after you do the task together once. Scripts are optional, and Claude can write those too when a step needs exact logic.

Related terms

AI AgentAn AI agent is software that uses an AI model to work toward a goal in a loop: it reads the situation, decides on an action, uses a tool such as a CRM, browser or email, checks the result and repeats until the job is done. Unlike a chatbot, it acts instead of only answering. Agentic WorkflowAn agentic workflow is a business process in which an AI agent handles some steps on its own: it plans the work, picks which tools to use, checks the result and tries again if it falls short. People define the goal, the limits and the approval points. Fixed automation, by contrast, follows the same steps every time. Context WindowA context window is the amount of text, measured in tokens, that an AI model can work with in a single request. It holds everything the model sees at once: instructions, conversation history, documents, tool results and its own answer. Anything outside the window is invisible to the model, and on the API, everything inside it is billed. GTM EngineeringGTM engineering is the practice of building the data and automation systems behind a company's go-to-market: finding target accounts, enriching contacts, running AI research, scoring and routing leads, and syncing clean records to the CRM and outreach tools. A GTM engineer builds these workflows, often in Clay, so sellers spend their time on conversations instead of spreadsheets.

Want these systems running inside your sales team?

Growth Cab is a go-to-market consulting firm for B2B tech companies selling $50,000+ deals. We build the outbound, LinkedIn and closing systems.

Apply to GC GTM Glossary