// Security Advisory Triage
mcp-sanitizer URL sanitizer allows localhost SSRF bypass
GitHub advisories — npm · published 2026-10-31 05:44 UTC
WHAT CHANGEDGitHub published GHSA-4mfg-r38w-w8fg on 28 October 2026, rating high severity. The mcp-sanitizer package URL sanitizer permits SSRF to localhost and IPv6 loopback addresses when an explicit port or bracketed IPv6 notation is used. Version 1.4.0 patches the vulnerability.
WHY IT MATTERSAn attacker can bypass URL sanitization to reach internal services that should be blocked, enabling server-side request forgery against localhost endpoints. This affects systems using mcp-sanitizer to validate URLs before fetching them, particularly in AI agent contexts where the Model Context Protocol processes external input.
WHAT TO DOCheck whether you run mcp-sanitizer below 1.4.0. If so, update to 1.4.0 in this cycle.
Model-reported confidence 85% (not measured accuracy)
Primary source ↗
written by anthropic/claude-sonnet-4.5
Show the reasoning — raw diff, triage verdict, confidence
RAW DIFF
+ghsa_id: GHSA-4mfg-r38w-w8fg +summary: mcp-sanitizer: URL sanitizer allows localhost/IPv6 loopback SSRF bypass via explicit port or bracketed IPv6 +severity: high +published_at: 2026-10-28T19:51:39Z +cve_id: +vulnerabilities[].package.name: mcp-sanitizer +vulnerabilities[].first_patched_version: 1.4.0
HOW IT WAS JUDGED
Triage signals below are model-reported assertions, preserved verbatim; they are not independent arithmetic verification.
| triage verdict | MATERIAL at 0.72, needed 0.5 |
|---|---|
| triage reasoning | SSRF in AI agent framework sanitizer affects services running MCP-based agents or tools. |
| signals | new advisory published · SSRF vulnerability in URL sanitizer · high severity rating · affects mcp-sanitizer package · patched version available (1.4.0) · MCP (Model Context Protocol) is AI agent framework component |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |