caption-clip
Warn
Audited by Snyk on Mar 4, 2026
Risk Level: MEDIUM
Full Analysis
MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).
- Third-party content exposure detected (high risk: 0.80). The skill explicitly downloads arbitrary public YouTube videos (SKILL.md Step 1), transcribes them via Deepgram, and then has Claude read and clean the resulting SRT (SKILL.md Step 5 / README "This step uses Claude's judgment"), so untrusted user-generated content from the open web is ingested and interpreted by the agent and could carry instructions that influence its behavior.
Audit Metadata