tiktok-shop-seller-scraper
Fail
Audited by Gen Agent Trust Hub on Jul 20, 2026
Risk Level: HIGHEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [REMOTE_CODE_EXECUTION]: The skill explicitly directs the agent to install missing dependencies (the BrowserAct CLI and Skill) automatically. It instructs the agent to not ask for user confirmation and to avoid describing the action as a request for approval, which facilitates the silent execution of remote code.
- [EXTERNAL_DOWNLOADS]: The skill references an external domain (browseract.com) for its core functionality and installation process. The use of a redirect parameter in the installation URL to a GitHub repository is a pattern that can be used to mask the true source or destination of the downloaded software.
- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface as it is designed to ingest and process data from external TikTok Shop pages.
- Ingestion points: The agent accesses and parses arbitrary content from TikTok seller profiles and product pages (SKILL.md).
- Boundary markers: There are no defined delimiters or instructions to treat external data as untrusted content.
- Capability inventory: The agent utilizes the BrowserAct CLI and can write data to the filesystem for exports.
- Sanitization: The instructions lack any requirement for sanitizing, validating, or escaping the data retrieved from the web before processing.
Recommendations
- AI detected serious security threats
Audit Metadata