voc-synthesis

Warn

Audited by Snyk on Mar 22, 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 SKILL.md explicitly ingests and processes untrusted, user-generated sources (see "Input Types" — "Review/comment mining" with examples like G2 reviews, LinkedIn comments, Reddit threads — and the Worked Example's "LinkedIn comment mining"), extracting verbatim quotes and using them to drive synthesis and downstream actions, so third‑party content can materially influence the agent's decisions.

Issues (1)

W011
MEDIUM

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

Audit Metadata
Risk Level
MEDIUM
Analyzed
Mar 22, 2026, 01:39 AM
Issues
1