tui-gap
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill defines a surface for indirect prompt injection by processing external input data (such as component properties and error messages) and interpolating it into structured output blocks for use in design tools. * Ingestion points: Data is accepted via the JSON Input Model containing fields like 'props' and 'error'. * Boundary markers: The skill uses specific headers (e.g., COMPONENT_SPEC, PENCIL_BATCH_DESIGN) to structure output, but does not define internal delimiters for the ingested content. * Capability inventory: The skill generates specifications for the 'Pencil MCP' tool but does not possess independent file-write or network capabilities. * Sanitization: There is no requirement or instruction for sanitizing input strings before they are used in the generated design specifications.
Audit Metadata