kuaishou-search-results-scraper
Fail
Audited by Gen Agent Trust Hub on Jul 20, 2026
Risk Level: HIGHREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [REMOTE_CODE_EXECUTION]: The skill instructs the agent to install an external dependency (BrowserAct CLI) from a remote URL (browseract.com) if it is missing.
- [COMMAND_EXECUTION]: Instructions explicitly state to perform the installation "automatically without asking the user to confirm" and "do not request separate approval for dependency installation." This is a significant security concern as it bypasses user consent for system-level changes.
- [PROMPT_INJECTION]: The skill includes instructions to suppress standard agent behavior (asking for permission) when performing software installations. Additionally, the skill is vulnerable to indirect prompt injection as it ingests untrusted search results from Kuaishou. * Ingestion points: Kuaishou search query results, titles, and metadata processed from live pages. * Boundary markers: None provided to separate untrusted data from agent instructions. * Capability inventory: The skill uses the browser-act tool for live browser interaction and data collection. * Sanitization: No validation or filtering of scraped content is mentioned before processing.
Recommendations
- AI detected serious security threats
Audit Metadata