The Agent Skills Directory

[Indirect Prompt Injection] (LOW): The skill is designed to process untrusted external text provided by users for copy editing. Ingestion points: External text and UI copy provided as input to the workflow. Boundary markers: None present in the instructions. Capability inventory: No risky capabilities such as file-write, network operations, or command execution were found in the skill. Sanitization: No sanitization or filtering logic is present.
[No Code] (INFO): The skill consists entirely of Markdown instructions. It does not include any scripts, binaries, or automated tasks, which inherently limits the potential attack surface to the reasoning layer only.

impeccable-clarify