The Agent Skills Directory

[COMMAND_EXECUTION]: The create-team.sh script executes administrative commands to manage the lifecycle of agent containers. This includes using docker exec and docker cp to inject configuration files directly into running containers, and mc mirror to synchronize team data with remote storage. It also invokes secondary orchestration scripts such as create-worker.sh and manage-teams-registry.sh to provision resources. The script handles administrative credentials from a secrets file to obtain necessary access tokens for Matrix room orchestration, which is documented as a standard practice for this platform.\n- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface as it interpolates user-supplied data, such as worker names and skill descriptions, into the instruction files (AGENTS.md and SOUL.md) of the created agents. \n
Ingestion points: User-provided arguments to create-team.sh including --workers, --worker-skills, and --worker-mcp-servers.\n
Boundary markers: The script utilizes  delimiters to separate injected coordination context from the rest of the agent's instructions.\n
Capability inventory: The skill possesses significant capabilities including the ability to execute docker commands, perform network requests to the internal Matrix server, and modify filesystem contents across agent directories.\n
Sanitization: The script uses jq to safely construct JSON payloads for API requests, while relying on template-based interpolation for Markdown documentation.

team-management