Restrict Google Discovery bundle URLs

Author: RhysSullivan

packages/plugins/google/src/sdk/discovery.test.ts

@@ -5,6 +5,8 @@ import { buildToolTypeScriptPreview } from "@executor-js/sdk/core";
 import {
   convertGoogleDiscoveryBundleToOpenApi,
   convertGoogleDiscoveryToOpenApi,
+  isGoogleDiscoveryUrl,
+  normalizeGoogleDiscoveryUrl,
 } from "./discovery";
 import { extract, parse } from "@executor-js/plugin-openapi";
 
@@ -44,6 +46,31 @@ const ConvertedSpec = Schema.Struct({
 
 const decodeConvertedSpec = Schema.decodeUnknownSync(Schema.fromJsonString(ConvertedSpec));
 
+it("accepts only supported HTTPS Google Discovery endpoints", () => {
+  expect(
+    normalizeGoogleDiscoveryUrl("https://www.googleapis.com/discovery/v1/apis/gmail/v1/rest/"),
+  ).toBe("https://www.googleapis.com/discovery/v1/apis/gmail/v1/rest");
+  expect(normalizeGoogleDiscoveryUrl("https://chat.googleapis.com/$discovery/rest?version=v1")).toBe(
+    "https://www.googleapis.com/discovery/v1/apis/chat/v1/rest",
+  );
+
+  expect(isGoogleDiscoveryUrl("https://www.googleapis.com/discovery/v1/apis/gmail/v1/rest")).toBe(
+    true,
+  );
+  expect(
+    isGoogleDiscoveryUrl("https://evilgoogleapis.com/discovery/v1/apis/gmail/v1/rest"),
+  ).toBe(false);
+  expect(isGoogleDiscoveryUrl("http://www.googleapis.com/discovery/v1/apis/gmail/v1/rest")).toBe(
+    false,
+  );
+  expect(
+    isGoogleDiscoveryUrl("https://www.googleapis.com/discovery/v1/apis/gmail/v1/rest?next=x"),
+  ).toBe(false);
+  expect(
+    isGoogleDiscoveryUrl("https://token@www.googleapis.com/discovery/v1/apis/gmail/v1/rest"),
+  ).toBe(false);
+});
+
 const normalizeOpenApiRefsForPreview = (node: unknown): unknown => {
   if (node == null || typeof node !== "object") return node;
   if (Array.isArray(node)) return node.map(normalizeOpenApiRefsForPreview);

packages/plugins/google/src/sdk/discovery.ts

@@ -231,38 +231,70 @@ const decodeDiscoveryMethod = Schema.decodeUnknownSync(DiscoveryMethod);
 const decodeDiscoveryResource = Schema.decodeUnknownSync(DiscoveryResource);
 const parseJson = Schema.decodeUnknownEffect(Schema.fromJsonString(Schema.Unknown));
 
-const normalizeDiscoveryUrl = (discoveryUrl: string): string => {
-  const trimmed = discoveryUrl.trim();
-  if (!URL.canParse(trimmed)) return trimmed;
-  const parsed = new URL(trimmed);
-  if (parsed.pathname !== "/$discovery/rest") return trimmed;
-  const version = parsed.searchParams.get("version")?.trim();
-  if (!version) return trimmed;
-  const host = parsed.hostname.toLowerCase();
-  if (!host.endsWith(".googleapis.com")) return trimmed;
+const DISCOVERY_SERVICE_PATH_RE =
+  /^\/discovery\/v1\/apis\/([A-Za-z0-9._-]+)\/([A-Za-z0-9._-]+)\/rest\/?$/;
+const DISCOVERY_VERSION_RE = /^[A-Za-z0-9._-]+$/;
+
+const serviceFromGoogleApisHost = (host: string): string | null => {
+  if (!host.endsWith(".googleapis.com")) return null;
   const rawService = host.slice(0, -".googleapis.com".length);
+  if (!rawService || rawService.includes(".")) return null;
   const service =
     rawService === "calendar-json"
       ? "calendar"
       : rawService.endsWith("-json")
         ? rawService.slice(0, -5)
         : rawService;
-  return service ? `${DISCOVERY_SERVICE_HOST}/${service}/${version}/rest` : trimmed;
+  return /^[a-z0-9][a-z0-9-]*$/.test(service) ? service : null;
 };
 
-export const isGoogleDiscoveryUrl = (url: string): boolean => {
-  const trimmed = url.trim();
-  if (!URL.canParse(trimmed)) return false;
+export const normalizeGoogleDiscoveryUrl = (discoveryUrl: string): string | null => {
+  const trimmed = discoveryUrl.trim();
+  if (!URL.canParse(trimmed)) return null;
   const parsed = new URL(trimmed);
+  if (parsed.protocol !== "https:" || parsed.username || parsed.password || parsed.hash) {
+    return null;
+  }
+
   const host = parsed.hostname.toLowerCase();
-  if (!host.endsWith("googleapis.com")) return false;
-  return parsed.pathname.includes("/discovery/") || parsed.pathname.includes("$discovery");
+  if (host === "www.googleapis.com") {
+    if (parsed.search) return null;
+    const match = parsed.pathname.match(DISCOVERY_SERVICE_PATH_RE);
+    const service = match?.[1];
+    const version = match?.[2];
+    return service && version ? `${DISCOVERY_SERVICE_HOST}/${service}/${version}/rest` : null;
+  }
+
+  const service = serviceFromGoogleApisHost(host);
+  if (!service || !["/$discovery/rest", "/$discovery/rest/"].includes(parsed.pathname)) {
+    return null;
+  }
+  const keys = [...parsed.searchParams.keys()];
+  const version = parsed.searchParams.get("version")?.trim();
+  if (keys.length !== 1 || keys[0] !== "version" || !version || !DISCOVERY_VERSION_RE.test(version)) {
+    return null;
+  }
+  return `${DISCOVERY_SERVICE_HOST}/${service}/${version}/rest`;
+};
+
+const normalizeDiscoveryUrl = (discoveryUrl: string): string => {
+  return normalizeGoogleDiscoveryUrl(discoveryUrl) ?? discoveryUrl.trim();
+};
+
+export const isGoogleDiscoveryUrl = (url: string): boolean => {
+  return normalizeGoogleDiscoveryUrl(url) !== null;
 };
 
 export const fetchGoogleDiscoveryDocument = Effect.fn("OpenApi.fetchGoogleDiscoveryDocument")(
   function* (discoveryUrl: string, credentials?: SpecFetchCredentials) {
+    const normalizedDiscoveryUrl = normalizeGoogleDiscoveryUrl(discoveryUrl);
+    if (!normalizedDiscoveryUrl) {
+      return yield* new OpenApiParseError({
+        message: "Google Discovery document URL must be a supported googleapis.com HTTPS Discovery endpoint",
+      });
+    }
     const client = yield* HttpClient.HttpClient;
-    const requestUrl = new URL(discoveryUrl);
+    const requestUrl = new URL(normalizedDiscoveryUrl);
     for (const [name, value] of Object.entries(credentials?.queryParams ?? {})) {
       requestUrl.searchParams.set(name, value);
     }

packages/plugins/google/src/sdk/plugin.test.ts

@@ -12,7 +12,7 @@
 // ---------------------------------------------------------------------------
 
 import { describe, expect, it } from "@effect/vitest";
-import { Effect, Layer } from "effect";
+import { Effect, Exit, Layer } from "effect";
 import { HttpClient, HttpClientRequest, HttpClientResponse } from "effect/unstable/http";
 
 import {
@@ -174,6 +174,43 @@ const bundlePlugins = () =>
   [googlePlugin({ httpClientLayer: discoveryHttpClientLayer }), memoryCredentialsPlugin()] as const;
 
 describe("Google bundle add flow", () => {
+  it.effect("rejects lookalike Discovery hosts before fetching bundle documents", () =>
+    Effect.scoped(
+      Effect.gen(function* () {
+        let requests = 0;
+        const blockedHttpClientLayer = Layer.succeed(HttpClient.HttpClient)(
+          HttpClient.make((request: HttpClientRequest.HttpClientRequest) =>
+            Effect.sync(() => {
+              requests += 1;
+              return HttpClientResponse.fromWeb(
+                request,
+                new Response("unexpected request", { status: 500 }),
+              );
+            }),
+          ),
+        );
+        const executor = yield* createExecutor(
+          makeTestConfig({
+            plugins: [
+              googlePlugin({ httpClientLayer: blockedHttpClientLayer }),
+              memoryCredentialsPlugin(),
+            ],
+          }),
+        );
+
+        const exit = yield* executor.google
+          .addBundle({
+            urls: ["https://evilgoogleapis.com/discovery/v1/apis/calendar/v3/rest"],
+            slug: "bad_google",
+          })
+          .pipe(Effect.exit);
+
+        expect(Exit.isFailure(exit)).toBe(true);
+        expect(requests).toBe(0);
+      }),
+    ),
+  );
+
   it.effect(
     "addBundle merges calendar+gmail+drive into one google integration with no tool-name collisions",
     () =>

packages/plugins/google/src/sdk/plugin.ts

@@ -33,7 +33,7 @@ import {
 import {
   convertGoogleDiscoveryBundleToOpenApi,
   fetchGoogleDiscoveryDocument,
-  isGoogleDiscoveryUrl,
+  normalizeGoogleDiscoveryUrl,
 } from "./discovery";
 import { decodeGoogleIntegrationConfig, type GoogleIntegrationConfig } from "./config";
 import { googleOpenApiBundlePreset } from "./presets";
@@ -83,7 +83,7 @@ const fetchGoogleBundleConversion = (
   ).pipe(Effect.flatMap((documents) => convertGoogleDiscoveryBundleToOpenApi({ documents })));
 
 const uniqueUrls = (urls: readonly string[]): readonly string[] => [
-  ...new Set(urls.map((url) => url.trim()).filter((url) => url.length > 0)),
+  ...new Set(urls.flatMap((url) => normalizeGoogleDiscoveryUrl(url) ?? [])),
 ];
 
 const describeGoogleAuthMethods = (record: IntegrationRecord): readonly AuthMethodDescriptor[] => {
@@ -326,13 +326,14 @@ export const googlePlugin = definePlugin((options?: GooglePluginOptions) => ({
   detect: ({ ctx, url }) =>
     Effect.gen(function* () {
       const trimmed = url.trim();
-      if (!trimmed || !isGoogleDiscoveryUrl(trimmed)) return null;
+      const discoveryUrl = normalizeGoogleDiscoveryUrl(trimmed);
+      if (!trimmed || !discoveryUrl) return null;
       const httpClientLayer = options?.httpClientLayer ?? ctx.httpClientLayer;
-      const conversion = yield* fetchGoogleDiscoveryDocument(trimmed).pipe(
+      const conversion = yield* fetchGoogleDiscoveryDocument(discoveryUrl).pipe(
         Effect.provide(httpClientLayer),
         Effect.flatMap((documentText) =>
           convertGoogleDiscoveryBundleToOpenApi({
-            documents: [{ discoveryUrl: trimmed, documentText }],
+            documents: [{ discoveryUrl, documentText }],
           }),
         ),
         Effect.catch(() => Effect.succeed(null)),
@@ -341,7 +342,7 @@ export const googlePlugin = definePlugin((options?: GooglePluginOptions) => ({
       return IntegrationDetectionResult.make({
         kind: "google",
         confidence: "high",
-        endpoint: trimmed,
+        endpoint: discoveryUrl,
         name: conversion.title,
         slug: DEFAULT_GOOGLE_SLUG,
       });

packages/plugins/google/src/sdk/presets.ts

@@ -1,3 +1,5 @@
+import { normalizeGoogleDiscoveryUrl } from "./discovery";
+
 export interface GooglePreset {
   readonly id: string;
   readonly name: string;
@@ -242,6 +244,8 @@ export const googleOAuthConsentScopesForPreset = (presetId: string): readonly st
 // ---------------------------------------------------------------------------
 
 const normalizeGooglePresetUrl = (url: string): string => {
+  const discoveryUrl = normalizeGoogleDiscoveryUrl(url);
+  if (discoveryUrl) return discoveryUrl;
   const trimmed = url.trim();
   if (!URL.canParse(trimmed)) return trimmed.replace(/\/$/, "");
   const parsed = new URL(trimmed);