prototype
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process design data from external sources, such as existing DESIGN.md files or user-provided reference sites. This creates a surface where instructions could be embedded in the design data.
- Ingestion points: DESIGN.md and user-supplied reference URLs (SKILL.md, Step 1).
- Boundary markers: None specified to delimit or ignore instructions within these sources.
- Capability inventory: The skill has the ability to write to DESIGN.md and create new HTML files.
- Sanitization: No explicit sanitization or filtering of external design data is mentioned.
- [COMMAND_EXECUTION]: The skill instructs the agent to create a self-contained HTML file outside of the repository's file system scope (SKILL.md, Step 4). This involves generating and writing files to runtime-determined paths.
Audit Metadata