The Agent Skills Directory

Indirect Prompt Injection (LOW): The skill is designed to process untrusted data from the local repository which may contain malicious instructions.
Ingestion points: The skill explicitly instructs the agent to search for and study external documentation and code files (e.g., using grep for 'design system', 'ui guide', etc.).
Boundary markers: There are no explicit instructions or delimiters provided to ensure the agent ignores potential instructions embedded within the design documents or code comments it analyzes.
Capability inventory: The skill instructions grant the agent the capability to read repository files and perform write/delete operations to align code with design standards and 'clean up' orphaned code.
Sanitization: No sanitization or validation logic is specified for the content retrieved from the repository files before it is processed by the agent.

impeccable-normalize