{"uuid": "1dbe0ed2-daf5-4e1a-857b-053e5bbcce49", "vulnerability_lookup_origin": "1a89b78e-f703-45f3-bb86-59eb712668bd", "author": "9f56dd64-161d-43a6-b9c3-555944290a09", "vulnerability": "CVE-2026-59973", "type": "seen", "source": "https://gist.github.com/alon710/8bd3e264f91a00f434cd5f27c8d28e71", "content": "# CVE-2026-59973: CVE-2026-59973: High-Severity Server-Side Request Forgery in FrontMCP and mcp-from-openapi\n\n&gt; **CVSS Score:** 8.5\n&gt; **Published:** 2026-09-11\n&gt; **Full Report:** https://cvereports.com/reports/CVE-2026-59973\n\n## Summary\nCVE-2026-59973 is a high-severity Server-Side Request Forgery (SSRF) vulnerability in FrontMCP and its underlying OpenAPI parsing library, mcp-from-openapi. The flaw allows authenticated attackers capable of importing or configuring OpenAPI specifications to bypass string-based hostname filtering mechanisms. By employing DNS wildcard loopbacks, HTTP redirects, or IPv4-mapped IPv6 address formatting, attackers can coerce the application into sending HTTP requests to internal networks, loopback adapters, and cloud metadata environments.\n\n## TL;DR\nA validation bypass in FrontMCP and mcp-from-openapi allows authenticated users to execute arbitrary Server-Side Request Forgery (SSRF) attacks against internal interfaces, bypassing existing string-based denylists through DNS wildcards, HTTP redirects, or IPv4-mapped IPv6 addresses.\n\n## Exploit Status: POC\n\n## Technical Details\n\n- **CWE ID**: CWE-918\n- **Attack Vector**: Network (AV:N)\n- **CVSS Score**: 8.5\n- **Exploit Status**: poc\n- **KEV Status**: Not Listed\n- **Impact Type**: Server-Side Request Forgery\n\n## Affected Systems\n\n- FrontMCP hosted tool generation services\n- Internal networks connected to FrontMCP containers\n- Cloud hosting platform metadata environments (IMDSv1/v2)\n- **frontmcp**: &gt;= 1.2.1 &lt; 1.5.0 (Fixed in: `1.5.0`)\n- **@frontmcp/adapters**: &gt;= 1.2.1 &lt; 1.5.0 (Fixed in: `1.5.0`)\n- **mcp-from-openapi**: &gt;= 2.3.0 &lt; 2.5.0 (Fixed in: `2.5.0`)\n\n## Mitigation\n\n- Upgrade FrontMCP to version &gt;= 1.5.0\n- Upgrade mcp-from-openapi to version &gt;= 2.5.0\n- Disable external schema resolution via empty allowedProtocols rules\n- Implement network egress restrictions on container environments hosting FrontMCP\n\n**Remediation Steps:**\n1. Review npm dependency tree for vulnerable versions of mcp-from-openapi (&lt; 2.5.0) and frontmcp (&lt; 1.5.0).\n2. Execute `npm install frontmcp@1.5.0` or update package.json to point to fixed versions.\n3. Locate initialization files containing `OpenapiAdapter.init` and ensure `refResolution.allowedProtocols` is empty or locked to verified hosts.\n4. Deploy egress firewalls or network security groups to restrict outward container traffic to internal RFC 1918 networks.\n\n## References\n\n- [FrontMCP Security Advisory](https://github.com/agentfront/frontmcp/security/advisories/GHSA-65h7-9wrw-629c)\n- [FrontMCP Release Version 1.5.0](https://github.com/agentfront/frontmcp/releases/tag/v1.5.0)\n- [mcp-from-openapi Release Version 2.5.0](https://github.com/agentfront/mcp-from-openapi/releases/tag/v2.5.0)\n\n\n---\n*Generated by [CVEReports](https://cvereports.com/reports/CVE-2026-59973) - Automated Vulnerability Intelligence*", "creation_timestamp": "2026-09-15T16:31:12.041885Z"}