The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill establishes an orchestration surface where a coordinator agent processes and evaluates outputs from multiple sub-agents (teammates).
Ingestion points: The skill ingests data from teammate status reports, command outputs, and shared project files ('ledgers') across multiple execution rounds.
Boundary markers: It explicitly implements a 'Gate' system requiring that the author and the verifier of a task be separate agents, providing a structural defense against accidental or malicious misinformation.
Capability inventory: The skill utilizes platform-native orchestration tools including spawn_teammate, send_message, interrupt_agent, create_goal, and schedule_create to manage the team lifecycle.
Sanitization: It mandates that all conclusions be supported by 'reproducible evidence' (specific commands or files) rather than relying on natural language assertions from teammates.
[SAFE]: The skill is primarily instructional and structural, defining best practices for agent-led project management. It does not contain hardcoded credentials, malicious persistence mechanisms, or unauthorized remote code execution patterns. All mentioned tools and workflows are consistent with standard agentic orchestration platforms.

strategic-coordinator