competitor-analysis

Warn

Audited by Snyk on Feb 20, 2026

Risk Level: MEDIUM
Full Analysis

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

  • Third-party content exposure detected (high risk: 0.90). The skill explicitly instructs its Phase 1 subagents to perform WebSearch and WebFetch of public sites (e.g., "WebFetch the product homepage" in Agent 1, site:twitter/x and social links in Agent 2, and "site:reddit.com" and "site:news.ycombinator.com" in Agent 7 of REFERENCE.md), meaning it ingests untrusted, user-generated third‑party content which the agent must read and use to drive subsequent synthesis and actions.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 20, 2026, 06:06 AM