feat(cli): send threatcrush/<version> user agent to our API

apps/cli/src/core/__tests__/user-agent.test.ts

@@ -0,0 +1,41 @@
+import { describe, expect, it } from "vitest";
+import { installUserAgent, isOurHost, USER_AGENT } from "../user-agent.js";
+import { PKG_VERSION } from "../version.js";
+
+function fakeFetch() {
+  const seen: { url: string; ua: string | null }[] = [];
+  const target = {
+    fetch: (async (input: Parameters<typeof fetch>[0], init?: RequestInit) => {
+      const url = typeof input === "string" ? input : input instanceof URL ? input.href : input.url;
+      seen.push({ url, ua: new Headers(init?.headers).get("user-agent") });
+      return new Response("ok");
+    }) as typeof fetch,
+  };
+  return { target, seen };
+}
+
+describe("user agent", () => {
+  it("names the CLI and its version in a form the nightly report parses", () => {
+    expect(USER_AGENT.startsWith(`threatcrush/${PKG_VERSION} `)).toBe(true);
+  });
+
+  it("only matches our own hosts", () => {
+    expect(isOurHost("https://threatcrush.com/api/ingest", "https://threatcrush.com")).toBe(true);
+    expect(isOurHost("https://supabase.threatcrush.com/rest/v1/x", "https://threatcrush.com")).toBe(true);
+    expect(isOurHost("http://localhost:3000/api/ingest", "http://localhost:3000")).toBe(true);
+    expect(isOurHost("https://evilthreatcrush.com/", "https://threatcrush.com")).toBe(false);
+    expect(isOurHost("https://feodotracker.abuse.ch/x", "https://threatcrush.com")).toBe(false);
+    expect(isOurHost("not a url", "https://threatcrush.com")).toBe(false);
+  });
+
+  it("adds the agent to our API, keeps other headers, leaves third parties and explicit agents alone", async () => {
+    const { target, seen } = fakeFetch();
+    installUserAgent(target);
+    installUserAgent(target); // idempotent: no double wrap
+    await target.fetch("https://threatcrush.com/api/ingest", { method: "POST", headers: { authorization: "Bearer x" } });
+    await target.fetch(new URL("https://threatcrush.com/api/auth/me"));
+    await target.fetch("https://discord.com/api/webhooks/1");
+    await target.fetch("https://threatcrush.com/api/x", { headers: { "User-Agent": "custom" } });
+    expect(seen.map((s) => s.ua)).toEqual([USER_AGENT, USER_AGENT, null, "custom"]);
+  });
+});

apps/cli/src/core/user-agent.ts

@@ -0,0 +1,47 @@
+import { apiUrl } from './cli-config.js';
+import { PKG_VERSION } from './version.js';
+
+// Node's fetch says "node" and nothing else, so threatcrush.com cannot tell a
+// CLI on 0.13.40 from one on 0.13.51, or the CLI from any other Node script.
+// The CLI and the daemon both call installUserAgent() once at startup; every
+// request to our own API that does not set a User-Agent then carries
+// `threatcrush/<version>`. Third-party hosts (abuse.ch, Discord, PagerDuty, a
+// scan target) are left alone: they get Node's default, as before.
+
+export const USER_AGENT = `threatcrush/${PKG_VERSION} (node ${process.version}; ${process.platform})`;
+
+/** True for the configured API host and threatcrush.com or any of its subdomains. */
+export function isOurHost(url: string, api = apiUrl()): boolean {
+  let host: string;
+  try {
+    host = new URL(url).hostname.toLowerCase();
+  } catch {
+    return false;
+  }
+  let apiHost = '';
+  try {
+    apiHost = new URL(api).hostname.toLowerCase();
+  } catch {}
+  return host === apiHost || host === 'threatcrush.com' || host.endsWith('.threatcrush.com');
+}
+
+function urlOf(input: Parameters<typeof fetch>[0]): string {
+  if (typeof input === 'string') return input;
+  if (input instanceof URL) return input.href;
+  return input.url;
+}
+
+/** Wraps `target.fetch` so requests to our API carry the CLI's user agent. Idempotent. */
+export function installUserAgent(target: { fetch: typeof fetch } = globalThis): void {
+  const base = target.fetch as typeof fetch & { __threatcrushUA?: true };
+  if (!base || base.__threatcrushUA) return;
+  const wrapped = ((input: Parameters<typeof fetch>[0], init?: RequestInit) => {
+    if (!isOurHost(urlOf(input))) return base(input, init);
+    const headers = new Headers(init?.headers ?? (input instanceof Request ? input.headers : undefined));
+    if (headers.has('user-agent')) return base(input, init);
+    headers.set('user-agent', USER_AGENT);
+    return base(input, { ...init, headers });
+  }) as typeof fetch & { __threatcrushUA?: true };
+  wrapped.__threatcrushUA = true;
+  target.fetch = wrapped;
+}

apps/cli/src/daemon-entry.ts

@@ -1,5 +1,8 @@
 #!/usr/bin/env node
 import { runDaemon } from './daemon/index.js';
+import { installUserAgent } from './core/user-agent.js';
+
+installUserAgent();
 
 runDaemon().catch((err) => {
   console.error('threatcrushd failed to start:', err);

apps/cli/src/index.ts

@@ -69,6 +69,9 @@ import {
 import { IpcClient } from "./core/ipc-client.js";
 import { findRunningDaemon } from "./daemon/pidfile.js";
 import { PKG_VERSION } from "./core/version.js";
+import { installUserAgent } from "./core/user-agent.js";
+
+installUserAgent();
 
 const LOGO = `
 ${chalk.green("  ████████╗██╗  ██╗██████╗ ███████╗ █████╗ ████████╗")}