The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to manage project tasks by executing a local Python script (./.trellis/scripts/task.py) through the shell. These commands are used for creating task directories and linking subtasks.
[EXTERNAL_DOWNLOADS]: To inform technical decisions, the skill leverages a research-first phase where sub-agents are spawned to perform web searches and fetch documentation from external sources.
[PROMPT_INJECTION]: The skill employs strong directive language such as 'Non-negotiable' and 'Core Principles' to enforce a specific workflow and interaction style. These are legitimate behavioral constraints for the AI agent's task-discovery persona and do not attempt to bypass safety filters.

trellis-brainstorm