The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to ingest and act upon untrusted data from external sources.
Ingestion points: The agent reads team feedback via the list_comments tool and processes arbitrary code from the local filesystem through the references parameter in tools like design_page and design_component.
Boundary markers: The instructions recommend prefixing code snippets with their file paths but do not specify robust delimiters or explicit instructions to ignore potentially malicious commands embedded within the data.
Capability inventory: The skill provides a wide range of project-modifying tools, including file writes (edit_page), theme modifications (edit_theme), and resource deletion (delete_component).
Sanitization: There is no mention of sanitization, filtering, or validation for the content retrieved from comments or code references before it is processed.
[EXTERNAL_DOWNLOADS]: The skill references and interacts with external resources provided by the vendor.
Evidence: It directs users to the Subframe documentation (docs.subframe.com) and involves uploading assets to Subframe's infrastructure using the create_asset_upload tool.

design