The Agent Skills Directory

[SAFE]: The skill facilitates the creation and organization of course materials using standard shell commands.
Evidence: Use of mkdir -p, git mv, and git commit for directory and file management in SKILL.md.
[COMMAND_EXECUTION]: The skill invokes a linting tool to ensure the created structure is valid.
Evidence: Execution of pnpm ai-hero-cli internal lint in SKILL.md.
[PROMPT_INJECTION]: The skill processes user-provided course plans, which represents an indirect prompt injection surface.
Ingestion points: Plan parsing and extraction of section/exercise names in SKILL.md.
Boundary markers: None present.
Capability inventory: Directory creation (mkdir), file movement (git mv), and command execution (pnpm) in SKILL.md.
Sanitization: The skill instructs the agent to enforce "dash-case (lowercase, hyphens)" for all generated names, which provides a layer of sanitization for user-provided strings before they are used in filesystem operations.
[DATA_EXPOSURE]: The skill does not access sensitive files or perform network operations.
Evidence: No detection of hardcoded credentials, sensitive system paths (e.g., .ssh, .env), or network utilities (e.g., curl, wget) in SKILL.md or agents/openai.yaml.

scaffold-exercises