customer-research

Warn

Audited by Snyk on Apr 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: 0.90). The skill's MY-ICP.md template explicitly instructs the agent to "Pull directly from interviews, Reddit posts, reviews, support tickets" (SKILL.md "Their Language" section), which requires ingesting untrusted, user-generated public content that will be used to drive decisions and downstream skills.

Issues (1)

W011
MEDIUM

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

Audit Metadata
Risk Level
MEDIUM
Analyzed
Apr 15, 2026, 01:35 PM
Issues
1