The Agent Skills Directory

[COMMAND_EXECUTION]: The skill requires the agent to execute shell commands to manage the file system and validate the structure, including mkdir, git mv, git commit, and pnpm ai-hero-cli internal lint.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to process external, untrusted "plans" provided by the user to drive the scaffolding process, creating a surface for injection attacks.
Ingestion points: The agent parses a text-based exercise plan (sections, exercise names, and variant types) from the user.
Boundary markers: Absent; no specific delimiters or instruction-ignore markers are defined to isolate the user-provided plan from the skill's logic.
Capability inventory: The skill possesses file system write capabilities (mkdir), version control operations (git), and the ability to execute a CLI tool via pnpm.
Sanitization: Absent; the instructions do not describe any validation, sanitization, or filtering of the plan content before it is used to generate directory paths or file content.

scaffold-exercises