The Agent Skills Directory

[SAFE]: The skill is entirely instructional, containing documentation, checklists, and code snippets intended to guide safe development practices. No executable scripts are included with the skill.- [PROMPT_INJECTION]: No patterns associated with prompt injection, safety bypasses, or instruction overrides were detected. The skill discusses prompt injection as a threat to be mitigated in user applications.- [DATA_EXFILTRATION]: No hardcoded credentials, sensitive file access, or unauthorized network operations were found. The skill follows best practices by demonstrating secret management through environment variables.- [REMOTE_CODE_EXECUTION]: The skill does not contain commands or patterns for downloading and executing remote code. It provides warnings against unsafe practices like using eval() or unverified shell execution.- [OBFUSCATION]: No Base64, zero-width characters, homoglyphs, or other obfuscation techniques were found in the skill content.

security-and-hardening