The Agent Skills Directory

[SAFE]: The skill provides robust security guidance for Node.js backends, including mandatory redaction of secrets in loggers, instructions to avoid leaking stack traces in error responses, and methods to prevent cross-request state leakage using AsyncLocalStorage.
[SAFE]: Evaluation files (such as evals/files/server.ts and evals/files/export.route.ts) contain common vulnerability patterns like path traversal, blocking event-loop calls (readFileSync), and unvalidated user input reaching database queries. These are explicitly used as test cases for the agent's auditing logic and are correctly identified as bugs/leaks by the skill's expected behavior definitions.
[SAFE]: The skill instructions emphasize secure practices such as schema-validated configuration, input/output validation at every boundary, and proper graceful shutdown procedures to prevent data loss or connection drops during deployment.

nodejs-backend