The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes local Python scripts and shell commands to manage development tasks and orchestrate sub-agents. It invokes scripts located in the .trellis/scripts/ directory, including get_context.py, plan.py, start.py, and task.py, to manage task directories and agent states.\n- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface. It ingests user-supplied feature requirements and descriptions, which are سپس interpolated into shell commands (e.g., in the plan.py call) and used to generate project documentation (prd.md) that guides the behavior of downstream agents.\n
Ingestion points: User-provided feature requirements and module involvements are captured via the agent's interaction flow as described in SKILL.md.\n
Boundary markers: Requirements are enclosed in double quotes when passed as command-line arguments and written to files using heredocs, providing basic delimiters but not a robust security boundary against adversarial content.\n
Capability inventory: The skill can execute shell commands, write to the filesystem, and initiate additional agent processes.\n
Sanitization: There is no evidence of input validation, filtering, or escaping of user-provided content before it is processed by orchestration scripts or written to task files.

parallel