bishwashere/HTTP

HTTP

Make a plain HTTP request from the daemon process — same machine the user is on. Use this whenever the user asks you to check a URL, hit an API, poll a JSON endpoint, or verify a service is up, including local URLs like http://localhost:3000/.... No browser needed.

Why this exists: the browse skill drives a real headless Chromium via Playwright. That is correct for rendered pages (login, JS SPA, screenshots) but wrong — and brittle — for plain JSON. Cron jobs that "check an API every N seconds" should use http, not browse.

Localhost works. The daemon runs on the user's own machine. URLs like http://localhost:3000/api/..., http://127.0.0.1:8080, http://192.168.1.x, or *.local are reachable. Do not refuse to hit them.

Commands

• get — Fetch a URL with GET. Set arguments.url. Returns { status, body, json?, durationMs, contentType, truncated }. Use this for "check this API", "is the server up", "what does this endpoint return".
• post — POST a body. Set arguments.url and arguments.body (string or object — object will be JSON-encoded with content-type: application/json).
• put / patch / delete — Same shape as post.
• head — Same as get but no body returned (use to check status only).

Optional on every action:

• arguments.headers — object of header name → value (e.g. { "Authorization": "Bearer <token>" })
• arguments.timeoutMs — request timeout in ms (default 10000, capped at 60000)

Result envelope

The result is JSON with these fields:

• ok — boolean (true when status is 2xx)
• status, statusText
• url, method, durationMs
• contentType
• body — text body, capped at 14000 chars
• truncated — true when the body was capped
• json — parsed body when content-type is JSON or the body parses as JSON

On failure (timeout, DNS, connection refused) the result is { "error": "...", "url", "method", "durationMs" }.

Tool schema

http_get
  description: Fetch a URL with GET. Use for any plain HTTP/JSON endpoint, including http://localhost:* on this machine. Returns { status, body, json?, durationMs }.
  parameters:
    url: string
    headers: object (optional)
    timeoutMs: number (optional)

http_post
  description: POST to a URL with an optional body and headers. Body may be a string or an object (object is JSON-encoded with content-type application/json by default).
  parameters:
    url: string
    body: string (optional)
    headers: object (optional)
    timeoutMs: number (optional)

http_head
  description: HEAD request — fetch only status and headers, no body. Use to check whether a URL is reachable / a service is up.
  parameters:
    url: string
    headers: object (optional)
    timeoutMs: number (optional)