web-app-security-audit

Pass

Audited by Gen Agent Trust Hub on Apr 18, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to use grep, glob, curl, and package auditing tools. These are used strictly for identifying vulnerabilities and checking security headers in the user-specified project.
  • [PROMPT_INJECTION]: The skill contains complex prompts for sub-agents to analyze specific security layers. These prompts are functional and do not attempt to override system safety constraints.
  • [SAFE]: The skill processes untrusted code files for auditing. While this represents a surface for indirect prompt injection, the skill's operational scope is limited to reading files and producing documentation, and the instructions provide a clear, deterministic workflow that reduces the risk of unintended behavior.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 18, 2026, 01:40 AM