Guide · Agent skill
How to use the API Connector Builder agent skill
Updated · written with AI from the project's own documentation and checked against our catalogue data
The api-connector-builder skill is a set of instructions that tells an AI agent how to add a new API connector or provider to a codebase that already has integrations. Instead of writing a generic HTTP client, the agent studies at least two existing integrations in your repo and copies their file layout, config style, auth approach, error handling, tests and registration steps. This guide explains what the skill does, how to install it, how to trigger it with plain requests, and what to check before you rely on its output.
API Connector Builderaffaan-m/ECC · MIT Download the skill Quick start
- 1Download the ZIP. Get the api-connector-builder ZIP from the page where you found this guide. It contains the SKILL.md file.
- 2Unzip into a skills folder. Unzip it into ~/.claude/skills/ for personal use, or into .claude/skills/ inside your project to share it with that repo.
- 3Or upload it in the Claude apps. In the Claude apps, open Settings, go to Capabilities, then Skills, and upload the skill.
- 4Make a request that names the service. Ask the agent to build a connector or provider for a named external service, such as Jira or Slack, following the existing integrations in your repo.
- 5Let the agent study existing integrations. The skill directs the agent to inspect at least two existing connectors first, so make sure those are in the repo you are working in.
- 6Review the new files. Compare the new layout, config, auth, error handling, tests and registration with the existing connectors the agent used as references.
- 7Run your own tests. The skill does not run anything. Run your project's test suite and test the connector against the real or a sandboxed service.
What is an agent skill, and what does this one do?
An agent skill is a folder containing a SKILL.md file with instructions that an AI agent can load when a task matches the skill's description. The agent reads those instructions and follows them while it works on your project.
This skill is for adding a native integration surface to a repository. Its stated goal is to match the host repository's patterns for connector layout, config schema, auth model, error handling, test style, and registration or discovery wiring. It is designed to avoid inventing a second integration architecture, and it tells the agent to start from existing in-repo connectors rather than from vendor documentation alone.
- Maps the house style by inspecting at least two existing connectors or providers (file layout, abstraction boundaries, config model, retry and pagination conventions, registry hooks, test fixtures and naming).
- Narrows the target integration to what the repo needs: auth flow, key entities, core read and write operations, pagination and rate limits, and webhook or polling model.
- Builds the connector in slices: config or schema, client or transport, mapping layer, connector or provider entry point, registration, and tests.
- Checks the result against a quality checklist, including config validation, explicit auth and error handling, and completed registry wiring.
Install the skill from the ZIP download
Download the ZIP for the api-connector-builder skill from the page where you found this guide. The skill ships as a single file, SKILL.md, so the ZIP contains that one file.
Unzip it into one of two locations depending on how you want to use it. A personal install in your user folder makes the skill available across projects. A project install puts it inside one repository so teammates who clone that repo get it too.
If you use the Claude apps instead of a local setup, you can upload the skill through the settings screen rather than copying files by hand.
- Personal install: unzip into ~/.claude/skills/ so the skill folder sits there.
- Project install: unzip into the .claude/skills/ folder inside your project.
- Claude apps: go to Settings, then Capabilities, then Skills, and upload the skill there.
~/.claude/skills/
.claude/skills/Trigger the skill with a natural request
You do not need a special command to use the skill. The agent decides to load it when your request matches the skill's description, which is about building a new connector or provider that must fit existing integrations. Ask for the outcome in plain language and name the external service.
The skill's own examples are written in Japanese, because the source file is in Japanese. In English, requests in the same spirit look like the ones below. Naming the existing code pattern helps the agent find the right reference files.
- Build a Jira connector for this project.
- Add a Slack provider that follows the existing provider pattern.
- Create a new integration for this API.
- Build a plugin that matches the connector style in this repository.
Check the output against the repo
Treat the agent's output as a draft to review, not a finished feature. The skill contains only guidance. It has no scripts and no network access of its own, so it will not run your tests, call the external API, or confirm that the connector works.
Use the checklist in the skill as your review list. Compare the new files with the existing connectors the agent cited, confirm that registration and discovery are wired up, and run your project's own test suite. If the repo expects docs or examples to change with a new connector, check that they were updated too.
- Confirm the new files follow the same layout as the existing connectors.
- Confirm config validation, auth, and error handling are explicit.
- Confirm the new connector is registered wherever the other connectors are.
- Run the project's existing tests and add tests in the repo's style.
- Update docs or examples if the repo normally does this for connectors.
Reference layouts the skill uses
The skill includes three example layouts, one for a provider-style Python package, one for a connector-style layout, and one for a TypeScript plugin. These are shapes to look for in your own repo. The agent is supposed to match your real layout, not force one of these templates onto your project.
If your repo uses a different structure, the skill's guard rails say to follow the repo. The skill also warns against copying an old connector pattern when the repo has a newer one.
providers/
existing_provider/
__init__.py
provider.py
config.py
integrations/
existing/
client.py
models.py
connector.py
src/integrations/
existing/
index.ts
client.ts
types.ts
test.tsSafety notes: scripts and network
The skill declares that it runs no scripts and uses no network. It is a single SKILL.md file of written guidance. Nothing in the skill itself executes code or contacts a server.
The connector the agent writes is another matter. Any integration code it produces may make network calls to the external service when you run it, and it may read credentials from config. Review the auth handling and where secrets are stored before you run or deploy the new code, and use your own repo's secret management.
- The skill itself: no scripts, no network use, nothing listed under requirements.
- The generated connector: it will contain whatever network and auth code the agent writes, so review it like any other code change.
- Credentials: keep API keys and tokens in your project's existing configuration and secret handling, not hard-coded in new files.
What people use it for
- Adding a Jira or Slack connector to an application that already has several third-party integrations.
- Adding a new provider to a plugin system so it uses the same config and registration pattern as the existing providers.
- Creating a TypeScript integration in a repo that keeps each integration in its own folder with index, client, types and test files.
- Getting a checklist for config, tests and registry wiring when adding the fifth or tenth integration to a codebase.
Frequently asked questions
Is the API Connector Builder skill free?
The source lists the skill under the MIT licence, which is a permissive open-source licence. Check the licence file in the source repository for the exact terms before you redistribute it.
What do I need to use this skill?
The skill lists no requirements. It runs no scripts and uses no network, so it is a single SKILL.md file. You do need an AI agent that supports skills and a codebase that already contains at least one or two integrations for the agent to copy.
Where do I install the api-connector-builder skill?
Unzip the downloaded ZIP into ~/.claude/skills/ for a personal install, or into .claude/skills/ inside a project for a shared install. In the Claude apps, you can instead upload it under Settings, then Capabilities, then Skills.
Does this skill work if my repo has no existing integrations?
Not well. The skill depends on existing in-repo examples to copy, and it gives little independent architecture guidance. For a project with no integration pattern, the skill's main method has nothing to match.
Should I use this skill or a generic HTTP client?
Use the skill when you want the new connector to look native to your existing code. For a quick one-off script that calls a single API, a generic HTTP client is simpler and the skill adds more than you need.
Will the skill test the connector for me?
No. The skill contains only guidance and runs no scripts. It tells the agent to add tests in the repo's style, but you still need to run your test suite and check the connector against the real service yourself.
New agent skills, repos and AI tools — every week
The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.
Our verdict
Use this skill if you maintain a codebase with several existing third-party integrations and want new connectors to match them closely. Pick something else if your project has no integration pattern yet, if you only need a single quick API call, or if you expect the skill to run or verify the connector automatically, since it only provides written guidance.
Download the skill