deep-research

Warn

Audited by Snyk on Mar 19, 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.md "Web Research" and "Ecosystem and Community Research" workflow explicitly directs the agent to search and read public, user-generated sources (e.g., Reddit, Hacker News, GitHub issues, app store reviews, ProductHunt) and use those findings to shape recommendations, so untrusted third-party content is ingested and can influence actions.

Issues (1)

W011
MEDIUM

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

Audit Metadata
Risk Level
MEDIUM
Analyzed
Mar 19, 2026, 05:31 PM
Issues
1