noon-search-results-scraper
Fail
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: HIGHEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill directs the agent to download the 'BrowserAct' CLI and related skills from a third-party website (
browseract.com). - [REMOTE_CODE_EXECUTION]: Instructions explicitly state that the agent should 'install the missing dependency automatically without asking the user to confirm' and 'do not request separate approval for dependency installation.' This pattern allows for the silent introduction and execution of arbitrary third-party code.
- [COMMAND_EXECUTION]: The skill requires the agent to execute system commands to run the BrowserAct CLI and handle the automated installation process.
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from web scraping, which could contain malicious instructions designed to influence the agent.
- Ingestion points: Search results, product metadata, and merchant info from Noon (SKILL.md).
- Boundary markers: No specific delimiters or instructions to ignore embedded commands are present.
- Capability inventory: The agent can execute CLI commands and perform network requests via the BrowserAct dependency (SKILL.md).
- Sanitization: There is no evidence of data validation or sanitization before the scraped content is returned to the user or agent context.
Recommendations
- AI detected serious security threats
Audit Metadata