express-attack-probe

Warn

Audited by Gen Agent Trust Hub on Jun 13, 2026

Risk Level: MEDIUMDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [DATA_EXFILTRATION]: The skill directs the agent to perform path traversal attacks to discover and read sensitive configuration files. Specifically, rule EXP-SF-001 involves requesting files such as .env and package.json using traversal sequences (e.g., ..%2F..%2F.env). This behavior represents a data exposure risk, despite being the stated purpose of the pentesting skill.
  • [PROMPT_INJECTION]: This skill presents an indirect prompt injection vulnerability surface.
  • Ingestion points: Untrusted data is ingested when the agent parses application source code (e.g., app.use, app.get) and crawls web routes to discover the application structure.
  • Boundary markers: The instructions lack explicit delimiters or guidance to distinguish between user-provided code/content and the skill's own operational instructions.
  • Capability inventory: The agent is empowered to perform various network operations, including POST requests with JSON payloads and header manipulation (e.g., spoofing X-Forwarded-For).
  • Sanitization: No sanitization or validation of the discovered source code or route content is performed before the agent processes it.
  • [SAFE]: The skill includes links to reputable external security documentation from OWASP, Snyk, and the official Express.js website. These resources are well-known and trusted in the developer community.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Jun 13, 2026, 12:45 PM
Security Audit — agent-trust-hub — express-attack-probe