The Agent Skills Directory

[COMMAND_EXECUTION]: The skill includes templates for generating shell commands like git commit and pytest. These are standard tools used for software versioning and testing and are used as intended within the development workflow.
[PROMPT_INJECTION]: The skill establishes a surface for indirect prompt injection because it processes user-provided requirements and transforms them into implementation plans that the agent is then instructed to follow using sub-skills. This risk is inherent to the function of a task-planning agent.
Ingestion points: User-provided feature specifications and requirements (as described in the Overview).
Boundary markers: The skill does not define specific delimiters or instructions to prevent the agent from obeying instructions embedded within the user's requirements.
Capability inventory: The skill directs the agent to modify the file system, execute development commands, and invoke other autonomous sub-skills like superpowers:executing-plans.
Sanitization: No sanitization or validation of the input specifications is performed before integration into the plan document.

writing-plans