The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the agent to install the agent-relay package globally using npm install -g agent-relay or run it via npx. This package is a core tool provided for the orchestration functionality described in the skill.
[COMMAND_EXECUTION]: The agent is guided to execute a variety of shell commands using the agent-relay CLI to manage infrastructure, spawn worker agents, and monitor their status. These operations are within the expected scope of a headless orchestrator.
[PROMPT_INJECTION]: The skill implements a multi-agent coordination pattern that creates an indirect prompt injection surface, where the lead agent processes input generated by workers. ● Ingestion points: Worker logs, direct messages, and channel communications are ingested via CLI commands like agent-relay agents:logs, agent-relay inbox, and agent-relay history (documented in SKILL.md). ● Boundary markers: No specific delimiters or instructions to ignore embedded commands within worker output are provided in the orchestration templates. ● Capability inventory: The orchestrator agent possesses shell execution capabilities and tool access required to manage the lifecycle of other agents as described in SKILL.md. ● Sanitization: The skill does not provide mechanisms for sanitizing or validating content received from worker agents before it is processed by the lead agent.

running-headless-orchestrator