customer-research

Warn

Audited by Snyk on Apr 30, 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's Mode 2 workflow and its referenced source guide explicitly instruct the agent to fetch and mine untrusted, user-generated content from public sites (e.g., Reddit, G2/Capterra, Hacker News, YouTube comments, LinkedIn, app store reviews) as part of its required research steps (SKILL.md and references/source-guides.md), so third-party content would be read and used to influence analysis and actions.

Issues (1)

W011
MEDIUM

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

Audit Metadata
Risk Level
MEDIUM
Analyzed
Apr 30, 2026, 04:39 PM
Issues
1