The Agent Skills Directory

[SAFE]: The skill is a collection of high-quality documentation and code examples designed to guide an AI agent in following professional software engineering standards. No security violations were detected.- [PROMPT_INJECTION]: The instructions establish a professional persona ("senior backend engineer") and set operational constraints. There are no attempts to bypass safety filters, override system instructions, or extract sensitive prompt data. The guidelines actually improve security posture by mandating strict validation (Zod) and explicit error handling.- [DATA_EXFILTRATION]: The skill demonstrates safe data handling practices. It specifically warns against using process.env directly for secrets, recommends a unifiedConfig pattern, and instructs that sensitive files like .env and config.ini should be gitignored to prevent accidental exposure.- [COMMAND_EXECUTION]: The skill references standard development tools and commands, such as npm test and local script execution (node scripts/test-auth-route.js). These are routine for development environments and are not used in a malicious context.- [EXTERNAL_DOWNLOADS]: The skill mentions well-known, industry-standard packages and services including Sentry (error tracking), Prisma (ORM), Zod (validation), and Express (framework). No downloads from untrusted, unknown, or suspicious remote sources are present.

backend-dev-guidelines