The Agent Skills Directory

[COMMAND_EXECUTION]: The skill is designed to perform command-line operations, including running tests and executing cdk deploy. These actions are controlled by the contents of a local tasks.md file and are restricted to the primary purpose of requirement implementation.
[INDIRECT_PROMPT_INJECTION]: The skill reads and processes the tasks.md file to identify and execute the next available task. While this represents a surface for indirect prompt injection if the file is tampered with by an external actor, the skill incorporates mandatory confirmation prompts ('ถามยืนยันก่อนเสมอ') which serve as a human-in-the-loop safety checkpoint.

req