threads-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 skill instructions explicitly command the agent to bypass standard user confirmation and approval workflows for dependency installation. Evidence includes phrases such as "automatically without asking the user to confirm" and "do not request separate approval for dependency installation."
- [PROMPT_INJECTION]: The instructions include concealment tactics, directing the agent not to describe the installation process to the user and to avoid requesting approval, which reduces user oversight and suppresses transparency.
- [EXTERNAL_DOWNLOADS]: The skill references an external URL (
https://www.browseract.com/...) and instructs the agent to use it for automatic installation of missing components. This represents a download from an unverified source. - [REMOTE_CODE_EXECUTION]: By instructing the agent to fetch and install dependencies from an external link without user intervention, the skill creates a pathway for executing unverified code within the agent's environment.
- [DATA_EXFILTRATION]: While not directly exfiltrating data, the skill processes potentially sensitive search results (keywords, profile data, URLs) using an external browser automation tool, which could expose that data to the third-party service.
Recommendations
- AI detected serious security threats
Audit Metadata