aindf-screen-author
Pass
Audited by Gen Agent Trust Hub on Oct 10, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [NO_CODE]: The skill contains only instructional documentation and JSON templates. No scripts, binaries, or third-party dependencies are included or referenced for installation.
- [DATA_EXFILTRATION]: Instructions in Rule 3 prohibit the generation of custom URLs or scripts, restricting output to plain text and validated design system enumerations.
- [INDIRECT_PROMPT_INJECTION]: The skill establishes a data ingestion surface by reading design system components via tools like get-component. Safety is addressed through mandatory use of the validate-screen tool to check all output against the design system schema and a prohibition on generating active content like scripts or handlers.
Audit Metadata