The Agent Skills Directory

[SAFE]: The skill serves as a management interface for other agent skills via the agentit tool. It does not contain any executable code, obfuscation, or data exfiltration patterns. It explicitly instructs the agent to perform integrity checks (SHA-256 validation) and avoid implicit trust in external content. The tool usage described (e.g., agentit skills activate, agentit skills candidates) is consistent with the skill's stated purpose of skill discovery and delivery. Furthermore, the skill provides defensive instructions such as failing closed if content is tampered with and maintaining strict isolation between worker contexts.

using-agent-skills