The Agent Skills Directory

[SAFE]: The skill is designed for administrative and development tasks related to AI agent orchestration. It includes defensive measures for handling 'runaway' agents, such as the slap protocol which prioritizes stabilization and draining of workflows before taking action.
[SAFE]: No malicious patterns, obfuscation, or unauthorized data exfiltration were detected. The skill uses standard engineering tools (e.g., git, shell, workflow APIs) in a manner consistent with its stated purpose of agent systems management.
[SAFE]: Although the skill processes untrusted data such as agent logs and transcripts (Category 8 risk), it emphasizes a human-in-the-loop model for any modifications and restricts sub-agents to non-overlapping tasks to prevent accidental or malicious interference with the host system.

agents