solo-swarm

Warn

Audited by Snyk on Feb 21, 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). This skill's SKILL.md explicitly instructs teammates to use MCP web_search/SearXNG, WebSearch, and WebFetch to search and scrape public sites (e.g., site:reddit.com, news.ycombinator.com, site:github.com, Product Hunt, G2, Capterra), so the agent will ingest untrusted, user-generated third‑party content that can influence its research and actions.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 21, 2026, 05:54 PM