The Agent Skills Directory

[SAFE]: The skill incorporates robust multi-tenancy security controls. It explicitly instructs the agent to reject user-supplied identifiers and instead rely on server-injected, HMAC-signed context to prevent Insecure Direct Object Reference (IDOR) attacks.
[SAFE]: References to external resources, such as the hive-mcp-client Node.js package and the hiveintelligence.xyz domain, are consistent with the developer's identity (hive-intel) and serve legitimate integration purposes.
[SAFE]: The skill includes clear 'Procedure' and 'Hand-off' sections that guide the agent toward safe and appropriate tool usage, ensuring that persistent monitoring is only established upon clear user intent.

hive-stateful-monitoring