websearch-deep

Warn

Audited by Snyk on Feb 15, 2026

Risk Level: MEDIUM
Full Analysis

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

  • Third-party content exposure detected (high risk: 1.00). The skill explicitly instructs the agent to use WebSearch and Fetch MCP to retrieve and synthesize open web content — including user-generated community sources (Stack Overflow, Reddit, GitHub discussions, public blogs) — so the agent will read and interpret untrusted third‑party material.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 15, 2026, 09:03 PM