testing-cors-misconfiguration

Pass

Audited by Gen Agent Trust Hub on Apr 7, 2026

Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADSDATA_EXFILTRATIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface where untrusted data enters the agent context through user-specified target URLs.
  • Ingestion points: Target URLs and endpoints provided as CLI arguments in scripts/agent.py.
  • Boundary markers: Absent.
  • Capability inventory: Network request operations via the requests library in scripts/agent.py.
  • Sanitization: Absent; user input is directly interpolated into request strings.
  • [EXTERNAL_DOWNLOADS]: Documentation in SKILL.md references 'corscanner', a well-known open-source tool for security professionals.
  • [DATA_EXFILTRATION]: The skill contains HTML and JavaScript templates that demonstrate how an attacker might exfiltrate data from a misconfigured server. These are provided as educational PoC examples and do not affect the user's data.
  • [COMMAND_EXECUTION]: Includes a Python script for auditing API endpoints and provides various bash examples for manual testing. These tools are used for their stated security-testing purpose.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 7, 2026, 12:40 AM