The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes shell commands such as mkdir -p to create nested directory structures, git commit and git mv for repository management, and pnpm ai-hero-cli internal lint to validate the created content. These operations are restricted to the local environment and the target project structure.
[INDIRECT_PROMPT_INJECTION]: The skill processes a user-defined plan to determine the naming and structure of new directories and files.
Ingestion points: Exercise plans provided by the user and parsed by the agent according to SKILL.md workflows.
Boundary markers: Not specified in the skill instructions.
Capability inventory: Filesystem directory creation, file writing (readme.md), and git state modification.
Sanitization: Content is validated by the internal ai-hero-cli linter for structural correctness (e.g., non-empty files, no broken links), but no specific sanitization of input strings is noted.

scaffold-exercises