The Agent Skills Directory

[SAFE]: The skill's instructions and extensive reference documentation focus exclusively on AI engineering best practices, evaluation frameworks, and safety alignment. It incorporates mature security concepts (e.g., OWASP LLM Top 10 2025) and mandates defensive design patterns like input sanitization and least privilege.
[SAFE]: There is no evidence of obfuscation techniques. All content is readable markdown, and no Base64 strings, zero-width characters, or suspicious Unicode tag characters were found.
[SAFE]: The skill does not include any executable scripts, shell commands, or unauthorized network operations. References to external systems (such as Anthropic API tool versions) are used purely for descriptive design guidance.
[SAFE]: No privilege escalation, persistence mechanisms, or hardcoded credentials were detected. The skill correctly identifies sensitive file paths (e.g., .aws/credentials, .ssh) as targets for redaction and safety filtering rather than extraction.

oracle