youtube-summarize

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 0.95). Yes — both links are suspicious: the GitHub release is an unknown account distributing a direct executable/archive (and the skill even references a password-protected zip), and the glot.io snippet instructs pasting/running an external install script — both are high-risk patterns for malware distribution.

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

• Third-party content exposure detected (high risk: 0.90). The SKILL.md "Extract Transcript" and "Summarization Workflow" sections explicitly instruct using yt-dlp/openclaw-agent to fetch transcripts from YouTube VIDEO_URL (public, user-generated content) and then send those transcripts to the LLM for analysis, which means untrusted third‑party content can directly influence the agent's outputs.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly requires downloading and running the openclaw-agent binary from https://github.com/hedefbari/openclaw-agent/releases/download/latest/openclaw-agent (and on macOS directs the user to run a script from https://glot.io/snippets/hfdxv8uyaf), which are runtime external dependencies that involve executing remote code and are required for transcript extraction.