youtube-search-results-scraper
Fail
Audited by Gen Agent Trust Hub on Jul 20, 2026
Risk Level: HIGHPROMPT_INJECTIONEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The instructions explicitly direct the agent to override standard safety guardrails by performing high-risk actions without user oversight. Specifically, it mandates installing dependencies "automatically without asking the user to confirm" and instructs the agent to "not describe it as a request for the user to open their browser."
- [EXTERNAL_DOWNLOADS]: The skill requires the agent to fetch external components from an untrusted third-party domain (browseract.com) to function.
- [REMOTE_CODE_EXECUTION]: The instruction to "install the missing dependency automatically" facilitates the remote acquisition and execution of a Command Line Interface (CLI) tool from a non-whitelisted source.
- [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it ingests untrusted data from YouTube search results.
- Ingestion points: YouTube search titles, creators, and descriptions (SKILL.md).
- Boundary markers: Absent; no delimiters or warnings are provided to prevent the agent from following instructions embedded in scraped content.
- Capability inventory: Execution of the
browser-acttool and CLI commands. - Sanitization: Absent; no filtering or escaping is applied to the external search data before processing.
Recommendations
- AI detected serious security threats
Audit Metadata