Functionize Skills: What Is Available and How to Install Them

Functionize publishes a library of ready-made skills — briefings that teach your AI tool how a particular application works, so it can write good tests for it straight away. There are skills for Salesforce, SAP, Workday, ServiceNow, Microsoft 365 and more, plus tooling skills for migrating an existing suite and for authoring a skill for your own bespoke application.

This page lists what is available, what each one adds, and how to install them.

Why these exist

An agent with no context has to discover your application by trial and error. It gets there, but it asks questions, makes reasonable-but-wrong assumptions, and takes several attempts.

A skill removes that. Ask for "a test that a service agent can resolve an incident" with the ServiceNow skill installed, and the agent already knows what an incident is, where it lives, and what resolving one looks like. The background is in Why Skills Matter.

The catalog

The one everybody needs

Application skills

Install the ones matching the systems you test.

Tooling skills

The migration and skill-builder add-ons are one-off tools, not day-to-day testing. Most teams have either a Playwright suite or a Selenium one, not both — install the matching migrator once, bring the suite across, and move on.

Installing them

These install into Claude Code as plugins. Run these one at a time:

/plugin marketplace add FunctionizeSandbox/claude_skills
/plugin install functionize-core@functionize

That is the baseline. Then add whichever match your systems:

/plugin install functionize-salesforce@functionize
/plugin install functionize-sap@functionize
/plugin install functionize-workday@functionize
/plugin install functionize-servicenow@functionize
/plugin install functionize-m365@functionize
/plugin install functionize-marketing-web@functionize

And the optional tooling:

/plugin install functionize-coverage@functionize
/plugin install functionize-playwright-migrate@functionize
/plugin install functionize-selenium-migrate@functionize
/plugin install functionize-skill-builder@functionize

That is it. You do not select a skill when you want one — Claude pulls in whichever is relevant as soon as you ask for something Functionize-related.

Getting access first

The skills library is currently private, so your GitHub account has to be granted access before the first command will work. If you are a Functionize customer, ask your Functionize contact or raise a ticket and we will add you — there is nothing for you to set up.

You also need to be signed in to GitHub on your machine. Check with gh auth status; if it says you are not logged in, run gh auth login. If that is unfamiliar, or the command is not found, whoever granted you access can do this one-time step with you.

Checking it worked

Run /plugin and look at the Installed list — functionize-core and anything else you chose should be there. Then just ask Claude to help with a Functionize test.

If you get a "marketplace not found" or an access error, it is almost always one of the two things above: not signed in to GitHub, or access not granted yet.

Testing something we do not have a skill for

Most applications are not Salesforce. If you test something in-house, industry-specific, or simply not on the list above, you have three options — and they are not mutually exclusive.

1. Write one yourself

A skill is a folder with a single text file in it, and writing one is not an engineering task. If you can explain your application to a new colleague, you can write one.

Our step-by-step guide has a complete, copyable example: How to Build a Skill in Claude for Your System Under Test. It covers the file format, what to include and what to leave out, how to install it, and how to hand it to business users so they can write tests in their own words.

Install functionize-skill-builder and Claude will help you author it, and look at functionize-claims-workbench for a worked example of the finished article.

2. Ask us to build it — at no charge

We build skills for customers' systems under test as part of onboarding, including bespoke in-house applications. You give us a walkthrough and access; we write the skill and hand it over for your team to own and extend. There is no cost attached.

See Complimentary Skill Building for Your System Under Test, or raise a ticket and ask.

3. Both

The pattern that works best: we build the first skill for your main system under test, your team sees how it is put together, and then you write the next ones yourselves. A skill is yours to extend — it should grow as your application does.

Skills work with any AI tool

The plugin commands above are the Claude Code route because it is the smoothest. But a skill is just a text file describing your application, so the same content works in Claude Desktop, and the underlying context helps whichever assistant you use.

The other half of the setup is connecting that tool to Functionize over MCP, so it can actually create and run tests rather than only talk about them. We have guides for Claude Code, Claude Desktop, Cursor, VS Code with Copilot, Gemini CLI, ChatGPT and others — start at Getting Started with the Functionize MCP Server.

The whole path, end to end

1. Get a skill for your application — install a ready-made one, write your own, or ask us to build it.
2. Load it into your AI tool — a plugin install, or a folder for a bespoke skill.
3. Connect that tool to Functionize over MCP, so it can create and run real tests. Work through Exercise 7 if you want a guided version.
4. Start testing — in your own words, in the editor where your work already lives.

If you are not sure which skills apply to your estate, ask your Functionize contact. Telling us what you test is usually a two-minute conversation and it saves installing things you do not need.