youtube-video-scraper
Fail
Audited by Gen Agent Trust Hub on Jul 20, 2026
Risk Level: HIGHEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill requires the installation of the BrowserAct CLI and Skill from an external URL (github.com/browser-act/skills) if they are not already present in the environment.
- [COMMAND_EXECUTION]: Instructions explicitly direct the agent to bypass human oversight during the installation process, stating the agent should install dependencies "automatically without asking the user to confirm" and should "not request separate approval."
- [REMOTE_CODE_EXECUTION]: The combination of automated external downloads and immediate execution of the downloaded CLI tool enables the execution of unverified code from a non-whitelisted source.
- [PROMPT_INJECTION]: The skill facilitates indirect prompt injection by scraping user-generated content (video descriptions, comments, and titles) from YouTube without implementing security controls.
- Ingestion points: YouTube metadata such as titles, descriptions, and comments are ingested into the agent context via the BrowserAct tool (SKILL.md).
- Boundary markers: Absent. The instructions do not specify any delimiters or warnings to ignore embedded instructions in the scraped data.
- Capability inventory: The agent is authorized to use the BrowserAct CLI to interact with the web and execute browser workflows.
- Sanitization: Absent. There is no mention of escaping, filtering, or validating the content retrieved from YouTube before the agent processes it.
Recommendations
- AI detected serious security threats
Audit Metadata