The Agent Skills Directory

[SAFE]: No security threats or malicious patterns were identified in the skill instructions or reference files. The skill operates purely within the conversational context to simulate user personas.
[SAFE]: The skill does not perform any network operations, data exfiltration, or access to sensitive credentials. It has no capability to interact with the external environment or local filesystem.
[SAFE]: No remote code execution or dynamic code evaluation patterns were found. The skill does not install or use external packages.
[SAFE]: The skill includes instructions to ingest data from other agents through structured handoff templates. While this presents a surface for indirect prompt injection, the risk is negligible because the skill lacks execution tools (such as shell or network access) to act on malicious instructions. Ingestion points: _AGENT_CONTEXT and references/handoffs.md; Boundary markers: Structured YAML templates; Capability inventory: No dangerous tools or system commands; Sanitization: Standard markdown formatting.

plea