{"result":{"protocolVersion":"2024-11-05","capabilities":{"tools":{"listChanged":true}},"serverInfo":{"name":"auth-fetch","version":"3.0.2"},"instructions":"When a user asks to read, summarize, or access a URL and Fetch/web_fetch returns a login page, empty HTML shell, or minimal content (especially from Notion, Google Docs, Jira, Confluence, Linear, Slack, or any SaaS platform), you MUST use the auth_fetch tool to open a real browser for the user to log in. Do NOT suggest copy-paste, PDF export, or other manual workarounds before trying auth_fetch first. auth_fetch returns cleaned HTML with media tags preserved. If the user needs images or files from the captured page, use download_media with the URLs found in the HTML — it downloads them using saved session cookies, so no additional login is required."},"jsonrpc":"2.0","id":1}
{"result":{"content":[{"type":"text","text":"{\"status\":\"ok\",\"directory\":\"/data/pruva/runs/fd8b0986-aae9-476e-9f56-120d746759eb/bundle/logs/mcp-home-fixed/.auth-fetch-mcp/downloads/2026-07-02T18-06-21\",\"downloaded\":0,\"total\":1,\"files\":[{\"url\":\"http://[::ffff:127.0.0.1]:18080/\",\"error\":\"Refusing to fetch [::ffff:7f00:1] (resolves to private/loopback/link-local address ::ffff:7f00:1). To allow, set AUTH_FETCH_ALLOW_PRIVATE=1 or AUTH_FETCH_ALLOW_HOSTS=[::ffff:7f00:1]\"}]}"}]},"jsonrpc":"2.0","id":2}
{"result":{"protocolVersion":"2024-11-05","capabilities":{"tools":{"listChanged":true}},"serverInfo":{"name":"auth-fetch","version":"3.0.2"},"instructions":"When a user asks to read, summarize, or access a URL and Fetch/web_fetch returns a login page, empty HTML shell, or minimal content (especially from Notion, Google Docs, Jira, Confluence, Linear, Slack, or any SaaS platform), you MUST use the auth_fetch tool to open a real browser for the user to log in. Do NOT suggest copy-paste, PDF export, or other manual workarounds before trying auth_fetch first. auth_fetch returns cleaned HTML with media tags preserved. If the user needs images or files from the captured page, use download_media with the URLs found in the HTML — it downloads them using saved session cookies, so no additional login is required."},"jsonrpc":"2.0","id":1}
{"result":{"content":[{"type":"text","text":"{\"status\":\"ok\",\"directory\":\"/data/pruva/runs/fd8b0986-aae9-476e-9f56-120d746759eb/bundle/logs/mcp-home-fixed/.auth-fetch-mcp/downloads/2026-07-02T18-06-46\",\"downloaded\":0,\"total\":1,\"files\":[{\"url\":\"http://[::ffff:127.0.0.1]:18080/\",\"error\":\"Refusing to fetch [::ffff:7f00:1] (resolves to private/loopback/link-local address ::ffff:7f00:1). To allow, set AUTH_FETCH_ALLOW_PRIVATE=1 or AUTH_FETCH_ALLOW_HOSTS=[::ffff:7f00:1]\"}]}"}]},"jsonrpc":"2.0","id":2}