The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands such as mkdir -p for directory creation, git commit and git mv for repository management, and pnpm ai-hero-cli internal lint to validate the created structure (SKILL.md).
[PROMPT_INJECTION]: The skill processes user-provided 'plans' to determine directory names and file content without explicit sanitization or boundary markers, creating a surface for indirect prompt injection (SKILL.md). * Ingestion points: The user-supplied 'plan' extracted in the workflow. * Boundary markers: Absent for the plan input. * Capability inventory: Shell command execution via mkdir, git, and pnpm (SKILL.md). * Sanitization: No validation is mentioned for directory names or readme content derived from the plan.

scaffold-exercises