design-md
Pass
Audited by Gen Agent Trust Hub on Sep 5, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill downloads and executes the
@google/design.mdpackage from the NPM registry usingnpx. This is the official tool for the open-source design specification. - [COMMAND_EXECUTION]: The skill provides instructions for the agent to execute shell commands for linting structure, comparing versions (
diff), and exporting tokens to Tailwind or DTCG formats. - [INDIRECT_PROMPT_INJECTION]: The skill ingests user-provided Markdown and YAML content as input for processing.
- Ingestion points: Reads existing
DESIGN.mdfiles or user-pasted design specs (referenced in SKILL.md workflow). - Boundary markers: None specific; the skill processes raw file content.
- Capability inventory: Includes file writing (
write_file) and shell execution (npx). - Sanitization: Standard CLI parsing by the
@google/design.mdtool is utilized, but no explicit agent-side sanitization of prompt data is specified.
Audit Metadata