cursor-api-key-management

Pass

Audited by Gen Agent Trust Hub on Mar 13, 2026

Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill requests broad shell access via the Bash tool to perform configuration checks and troubleshooting.\n
  • Evidence: SKILL.md specifies allowed-tools: "Read, Write, Edit, Bash(cmd:*)".\n
  • Evidence: references/troubleshooting.md includes example curl commands for API testing.\n- [DATA_EXFILTRATION]: The skill is designed to read from and write to sensitive configuration files that contain authentication secrets.\n
  • Evidence: references/configuration-methods.md references ~/.config/Cursor/User/settings.json, ~/.zshrc, and ~/.bashrc as primary configuration targets.\n
  • Evidence: references/security-best-practices.md provides commands to inspect and modify permissions on these sensitive files.\n- [PROMPT_INJECTION]: The skill establishes a high-capability environment that interacts with sensitive data files, creating a surface for indirect prompt injection.\n
  • Ingestion points: User configuration files and shell profiles (references/configuration-methods.md).\n
  • Boundary markers: None present to distinguish between legitimate configuration and embedded instructions.\n
  • Capability inventory: Full file manipulation (Read, Write, Edit) and shell execution (Bash) capabilities (SKILL.md).\n
  • Sanitization: No evidence of input validation or escaping for data retrieved from these files.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 13, 2026, 11:18 AM