The Agent Skills Directory

[SAFE]: The skill is entirely instructional and provides conceptual code snippets for internal agent communication using hypothetical libraries.
[PROMPT_INJECTION]: No malicious instructions, jailbreak attempts, or safety filter bypasses were identified in the documentation or metadata.
[DATA_EXFILTRATION]: No network operations (curl, wget, fetch) or access to sensitive system paths (like credentials or SSH keys) were found in the code examples.
[COMMAND_EXECUTION]: No shell command execution, subprocess spawning, or dynamic code evaluation (eval/exec) patterns were detected.

Group Collaboration