system-connector

Build deterministic Node.js connectors to third-party APIs and systems.

Updated Aug 6, 2026
One-click install
npx skills add https://github.com/ferrarifankid04/ai-skills-public --skill system-connector-ferrarifankid04
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: system-connector
Source: https://github.com/ferrarifankid04/ai-skills-public/tree/main/claude-code/skills/system-connector
Command: npx skills add https://github.com/ferrarifankid04/ai-skills-public --skill system-connector-ferrarifankid04

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Connecting an AI assistant to a third-party system, SaaS product, or internal tool usually requires manual API research, auth setup, and custom scripting. This Skill automates that entire process, producing a working, tested connector instead of fragile ad-hoc snippets. ## Core Features & Use Cases - Guided six-phase workflow: Runs interview, discovery, research, build, verify, and install phases to produce a deterministic connector rather than guessing at runtime. - Multiple output types: Produces an MCP config snippet, a SKILL.md plus stdlib-only Node.js helper script with self-tests, a full MCP server, or an honest stop when integration is not feasible. - Secure credential handling: Stores secrets in a local .env file the assistant never reads, with per-variable acquisition guidance and a privacy guarantee. - Use Case: A user says "connect me to our regional ERP system." The Skill checks for existing MCP servers, researches the vendor's API docs, builds a Node.js helper covering all read endpoints with tests, installs it, and walks the user through filling in credentials. ## Quick Start Ask the assistant to connect you to any external service by name, for example "set up an integration with my HubSpot account."

Frequently Asked Questions about system-connector

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I connect an AI assistant to a third-party API?

Name the service you want to connect to, and the Skill checks for existing MCP servers or skills first, then researches the official API docs and builds a Node.js helper script with tests. You only fill in credentials in a local .env file.

What does the system-connector skill produce as output?

Depending on the target system, it produces an MCP config snippet, a SKILL.md bundle with a stdlib-only Node.js helper script and self-tests, a full MCP server, or an honest explanation when integration is not feasible.

Does the helper script require npm install or external dependencies?

No. The generated helper targets Node.js 18 or later using only built-in modules like global fetch, node:fs, and node:test. There is no package.json, no npm install, and no third-party dependencies.

How are API credentials handled and stored?

Credentials live in a local .env file at <host-root>/connectors/<system>/.env that only the helper script reads. The assistant never asks for, reads, or logs credential values in chat, and setup is idempotent so existing values are never overwritten.

Can the connector perform write operations like creating records?

By default the helper covers only non-destructive read, list, query, and search endpoints. Write operations are added later only when you explicitly request a specific write action, with confirmation before the first real call.

What happens if the target system has no public API?

The Skill falls back to non-API integrations such as CSV/Excel import-export, local database reads, or desktop UI automation. If no usable integration surface exists at all, it stops honestly rather than shipping something that will not work.