web-performance
Fail
Audited by Gen Agent Trust Hub on Feb 15, 2026
Risk Level: HIGHCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION] (HIGH): The script 'scripts/performance-auditor.py' executes the 'curl' utility via 'subprocess.run' to measure page load times. This allows the agent to make network requests to arbitrary URLs, which could be exploited for Server-Side Request Forgery (SSRF) or to access internal network resources.
- [PROMPT_INJECTION] (HIGH): The skill possesses a significant Indirect Prompt Injection surface. It is designed to ingest and analyze external web content (via URLs). A malicious webpage could contain instructions that attempt to hijack the agent's logic once the content is processed. Ingestion point: 'url' parameter in 'scripts/performance-auditor.py'. Capability: Network interaction via 'curl' and local file reading. Sanitization: Absent.
- [EXTERNAL_DOWNLOADS] (LOW): 'README.md' contains instructions for installing the 'lighthouse' package from npm. This is a common utility but represents an external dependency.
Recommendations
- AI detected serious security threats
Audit Metadata