Video Summarizer

Fail

Audited by Snyk on Feb 24, 2026

Risk Level: HIGH
Full Analysis

HIGH W007: Insecure credential handling detected in skill instructions.

  • Insecure credential handling detected (high risk: 0.80). The skill usually reads API keys from .env (safe), but it explicitly shows and documents passing an API key as a literal CLI argument (e.g., --api-key "sk-or-v1-YOUR_KEY"), which encourages embedding secrets verbatim in generated commands and thus creates an exfiltration risk.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

  • Third-party content exposure detected (high risk: 0.90). The skill explicitly ingests and processes user-generated public content (e.g., YouTube video URLs in SKILL.md Quick Start and CLI examples, and social platforms listed in README like Instagram/TikTok/Twitter/X/Reddit and Google Drive/Dropbox), and the agent is required to read and summarize those third‑party transcripts as part of its workflow, which could enable indirect prompt injection.
Audit Metadata
Risk Level
HIGH
Analyzed
Feb 24, 2026, 01:17 AM