SearXNG Basic Authentication Credentials Exposed Through MCP Logs and JSON-RPC Error Responses
GHSA-hjwh-xvfw-qrwj · published 2026-08-19
- Severity
- Medium
- CVSS score
- 5.5
- Chance of exploitation (EPSS)
- No CVE
- Fixed in
- 1.12.0
This is a sensitive credential disclosure vulnerability. The following parties may obtain the credentials: 1. A connected MCP client receiving logging notifications. 2. A client capable of invoking a tool and receiving JSON-RPC errors. 3. A user or process with access to captured stderr output. 4. A centralized logging or monitoring system collecting application logs. 5. Other users with access to shared log files or container logs. The exposed credentials may allow an attacker to authenticate directly to the configured SearXNG instance. Depending on the SearXNG deployment and the permissions associated with the account, this may allow: 1. Unauthorized use of a private SearXNG service. 2.
Upgrade mcp-searxng to 1.12.0 or later. If you start it from an MCP config with npx or uvx, pin that version so an old cached copy is not used.
Apply redactSearxngInstanceUrl() before including any SearXNG URL in console or MCP logging: Do not include raw configuration values in validation errors. A generic error can be returned instead: For malformed URLs: The following additional changes are recommended: 1. Redact URLs before writing them to stderr. 2. Redact secrets before sending MCP logging notifications. 3. Avoid including raw environment-variable values in exceptions. 4. Avoid returning detailed stack traces containing secrets to MCP clients. 5. Mark SEARXNG_URL as secret in .mcp/server.json: 6.
| Package | Registry | Affected | Fixed in |
|---|---|---|---|
| mcp-searxng | npm | < 1.12.0 | 1.12.0 |
- CriticalPhoenix: Format String Injection in applyChatTemplatearize-phoenix
- CriticalFlowise: Unauthenticated OAuth2 token refresh endpoint returns access tokens — enables token theft for any connected serviceflowise
- CriticalMLflow: Environment variable injection in AI Gateway secrets enables server-side credential exfiltrationmlflow
- CriticalvLLM has RCE In Video Processingvllm
- HighConfig revision hashes could expose password verifiersopenclaw
- Highn8n: Shared Credential Header Leak via HTTP Request Pagination Expressionn8n, n8n, n8n
- HighSearXNG MCP Server: DNS-resolved Private Hostname SSRF in `web_url_read`mcp-searxng
- HighSearXNG MCP Server: Unbounded Response Body Read Bypasses URL Size Limit in `web_url_read`mcp-searxng
- MediumSearXNG MCP Server is Vulnerable to SSRF in web_url_read: the internal-address guard is disabled by default (MCP_HTTP_HARDEN off)mcp-searxng
- MediumSearXNG MCP Server: Additional hardened-mode SSRF bypassesmcp-searxng
The impact, patch and workaround text comes from the advisory, shortened (GitHub Advisory Database, CC BY 4.0). Proof-of-concept and reproduction sections and code are left out; the original advisory has them.