The Agent Skills Directory

[SAFE]: The skill establishes a robust pre-execution risk classification system for AI agents, enhancing operational safety.
[SAFE]: It correctly identifies high-risk targets such as sensitive configuration files (.env, credentials) and destructive operations (git force-push, database deletions), requiring higher levels of justification before execution.
[SAFE]: No malicious patterns, such as prompt injection, data exfiltration, obfuscation, or unauthorized remote code execution, were identified in the instructions.

intent-recognition