The Agent Skills Directory

[COMMAND_EXECUTION]: The skill makes extensive use of the alera CLI tool to manage system tasks and terminals. Evidence includes commands like alera terminal write, which allows the agent to send text and control input to specific terminal handles, and alera orchestration agent-spawn, which initiates new agent processes.
[DYNAMIC_EXECUTION]: The orchestration workflow involves reading agent profiles that contain specific commands to be executed. The agent-spawn primitive resolves these commands host-side based on the selected profile (e.g., agent-spawn --profile <name>).
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from external sources and other agents, creating a potential attack surface.
Ingestion points: The messaging system (alera orchestration inbox) can receive banners and questions from external callers (ext:...), and the task context (alera orchestration context) contains specifications for worker tasks.
Boundary markers: The instructions specify that workers should follow a "dispatch preamble" which installs a terminal-scoped context, and emphasize responding only through the designated commands in external banners.
Capability inventory: The skill has the capability to write to terminals, spawn new agents, and transfer coordinator ownership across the Alera environment.
Sanitization: The skill instructions rely on the Alera runtime host to enforce lifecycle authority and audited recovery paths, rather than explicit string sanitization within the instructions themselves.

alera-orchestration