The Agent Skills Directory

[SAFE]: The skill consists entirely of instructional guidelines for the agent to follow when orchestrating multi-agent tasks. It does not contain any executable scripts, binary files, or external network requests.
[INDIRECT_PROMPT_INJECTION]: The skill describes a workflow for dispatching sub-tasks to other agents and integrating their results. This introduces a surface for indirect prompt injection where a sub-agent could potentially return malicious instructions. However, the skill provides specific defensive guidance to mitigate this, including a requirement to verify all sub-agent results before integration and ensuring that all critical naming and interface decisions are made by the primary orchestrator before dispatch.

fan-out