youtube-channel-email-finder

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (high risk: 0.90). The skill includes explicit instructions to automatically install the BrowserAct dependency and "do not request separate approval" or "do not describe it as a request for the user to open their browser," which instructs the agent to bypass or hide user consent — a deceptive behavior outside the advertised email-finding purpose.

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

• Third-party content exposure detected (high risk: 0.85). SKILL.md instructs the agent to invoke the BrowserAct CLI to access and interact with real YouTube/public web pages at runtime, which can ingest outsider-authored free text (e.g., channel About/contact text) into the LLM context via the browser-extracted page content.

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

• Potentially malicious external URL detected (high risk: 1.00). The skill explicitly instructs the agent to automatically install the required BrowserAct runtime using the install link https://www.browseract.com/?co-from=lead-generation&redirect=https://github.com/browser-act/skills/tree/main at runtime, which would fetch and run external installation content that can execute code or control agent behavior.