The Agent Skills Directory

[DATA_EXPOSURE]: The skill instructions direct the agent to read local project documentation such as README, AGENTS.md, and existing architecture documents to orient itself. This is a standard and necessary function for a design brainstorming tool.
[COMMAND_EXECUTION]: The skill performs file system operations specifically for reading codebase implementation details and writing the final design specification to a new file in docs/specs/. These actions are restricted to the expected output path and are consistent with the skill's primary purpose.
[PROMPT_INJECTION]: The skill defines a specific conversational style (e.g., "one question at a time", "YAGNI check") and scope control rules. These are behavioral guidelines for the AI to follow during a workshop and do not attempt to bypass core safety filters or override system-level instructions.

spec-brainstorm