The Agent Skills Directory

[SAFE]: The skill provides legitimate architectural and security guidance for backend development. It explicitly encourages the use of input validation schemas (e.g., Zod, Pydantic) to prevent mass assignment and other injection vulnerabilities.
[DATA_EXPOSURE]: No evidence of hardcoded credentials or sensitive data exposure was found. The instructions emphasize proper error handling to ensure internal stack traces are never leaked to the client.
[REMOTE_CODE_EXECUTION]: No remote code execution patterns, dynamic code generation, or suspicious package installations are present. The skill recommends standard, well-known libraries for its stated purpose.
[PROMPT_INJECTION]: The content does not contain instructions to bypass safety filters or override agent constraints. It uses natural instructional language focused on engineering standards.
[COMMAND_EXECUTION]: No dangerous shell commands or privilege escalation attempts were detected. The only execution mentioned is targeted testing (e.g., npm test), which is standard practice in development environments.

backend