The Agent Skills Directory

[COMMAND_EXECUTION]: Employs mkdir, git mv, and pnpm ai-hero-cli internal lint to manage and validate the exercise directory structure. These commands are local and consistent with the skill's documented purpose.
[PROMPT_INJECTION]: Ingests user-supplied exercise plans to determine folder names and hierarchy (Indirect Prompt Injection surface).
Ingestion points: Processes exercise plans extracted from user input as described in the Workflow section of SKILL.md.
Boundary markers: No specific delimiters are used for the plan data.
Capability inventory: Performs directory creation (mkdir), file moves (git mv), and file writes (readme.md) as detailed in SKILL.md.
Sanitization: Enforces naming conventions (dash-case) and uses an internal linter to verify the resulting structure.

scaffold-exercises