diff --git a/src/crawl.mjs b/src/crawl.mjs
index ce9dec3..0881fdc 100644
--- a/src/crawl.mjs
+++ b/src/crawl.mjs
@@ -16,7 +16,9 @@ export function createCrawlGate(env = process.env) {
     coinpay: { apiKey },
     payTo,
     contact: 'mailto:anthony@profullstack.com',
-    openPaths: ['/healthz', '/llms.txt', '/install.sh', '/install', '/commands.json', '/blog/rss.xml', '/blog/feed.xml'],
+    openPaths: ['/healthz', '/llms.txt', '/install.sh', '/install', '/commands.json', '/blog/rss.xml', '/blog/feed.xml',
+      // The legal pages: a connector directory's review crawler must read them.
+      '/privacy', '/terms'],
   });
 }
 
diff --git a/test/crawl.test.mjs b/test/crawl.test.mjs
index 1b7ae4b..86ddd10 100644
--- a/test/crawl.test.mjs
+++ b/test/crawl.test.mjs
@@ -44,3 +44,11 @@ test('a browser and the health check are not charged', async () => {
   assert.equal((await get('/', CHROME)).status, 200);
   assert.equal((await get('/healthz', { 'user-agent': GPT })).status, 200);
 });
+
+test('legal pages are open to crawlers and browsers alike', async () => {
+  const CLAUDEBOT = 'Mozilla/5.0 (compatible; ClaudeBot/1.0; +claudebot@anthropic.com)';
+  for (const path of ['/privacy', '/terms']) {
+    for (const ua of [GPT, CLAUDEBOT]) assert.equal((await get(path, { 'user-agent': ua })).status, 200, `${path} ${ua}`);
+    assert.equal((await get(path, CHROME)).status, 200);
+  }
+});