The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses bun run to invoke local scripts and echo to append to a JSONL log file for execution tracking. These are standard operations for tool invocation and logging.
[COMMAND_EXECUTION]: The skill mentions and provides examples of using TeamCreate, TaskCreate, and SendMessage tools, which are part of the platform's orchestration layer.
[SAFE]: The skill does not perform any sensitive data access, exfiltration, or unauthorized network operations.
[SAFE]: Instructions for using isolation: "worktree" represent a security best practice for parallel file edits to prevent conflicts and ensure clean environments.

Delegation