The Agent Skills Directory

[SAFE]: The skill serves as a procedural guide for agent orchestration and task management. It does not implement executable logic, perform network operations, or access sensitive local files.
[SAFE]: No evidence of prompt injection, obfuscation, or persistence mechanisms was found. The instructions regarding environment variables such as SHIPYARD_TEAMS_ENABLED and CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS refer to standard platform feature flags rather than sensitive credentials or data exfiltration.
[SAFE]: The code snippets provided are illustrative examples of tool usage (Task()) and lifecycle management (Team creation/deletion) within a specific agent framework. They do not involve remote code execution from untrusted sources or privilege escalation.

parallel-dispatch