@x12i/connectix-tools
v0.1.1
Published
Machine-usable AI tool definitions and runners for the four curated Connectix operations (connect.listProviders, connect.searchActions, connect.getAction, connect.executeAction).
Downloads
274
Maintainers
Readme
@x12i/connectix-tools
Machine-usable AI tool definitions and runners for the four curated Connectix operations.
Callers (AI Skills, FuncX, Memorix Pipeline Services) expose these as LLM tools. They stay executor-source agnostic: no OpenConnector-, n8n-, or Nango-specific logic ever reaches a caller.
npm install @x12i/connectix-toolsTool definitions
import { CONNECT_TOOLS } from "@x12i/connectix-tools";
// [{ name: "connect_list_providers", operationId: "connect.listProviders", inputSchema, ... }, ...]| Tool name | Operation | Purpose |
|-----------|-----------|---------|
| connect_list_providers | connect.listProviders | List curated providers |
| connect_search_actions | connect.searchActions | Search curated actions |
| connect_get_action | connect.getAction | Fetch one action descriptor |
| connect_execute_action | connect.executeAction | Run an action (stub or live) |
Every definition carries JSON Schema inputSchema/outputSchema suitable for LLM tool registries.
Run against a live service
import { createHttpConnectToolRunner } from "@x12i/connectix-tools";
const runner = createHttpConnectToolRunner({
baseUrl: "http://127.0.0.1:9400",
callerId: "service:ai-skills",
callerToken: process.env.CONNECTIX_AI_SKILLS_TOKEN ?? "change-me-caller",
});
await runner.run({
tool: "connect_execute_action",
input: { actionId: "example.echo", bindingRef: "demo-echo", input: { message: "hi" }, stubEnabled: true },
});Run against the package simulator (no service, no network)
import { createConnectToolRunner } from "@x12i/connectix-tools";
import { createSimulatorPipelineCaller } from "@x12i/connectix-simulator";
const runner = createConnectToolRunner(createSimulatorPipelineCaller());License: UNLICENSED (X12I internal; published for X12I service consumers).
