awesome-design-md

Warn

Audited by Gen Agent Trust Hub on Aug 19, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [UNVERIFIABLE_DEPENDENCIES_AND_REMOTE_CODE_EXECUTION]: The skill uses npx --yes getdesign@latest to execute remote code from the npm registry. This pattern downloads and runs the latest version of the getdesign package from a third-party vendor without version pinning or integrity validation, allowing for potential supply chain attacks.
  • [INDIRECT_PROMPT_INJECTION]: The skill establishes a workflow where an agent ingests untrusted data (design templates) and is explicitly instructed to follow them as the "source of truth." This creates a vulnerability where malicious instructions embedded in the design files could influence the agent's behavior during UI implementation tasks.
  • Ingestion points: Reads the contents of DESIGN.md or files within <slug>/DESIGN.md generated by the external CLI tool.
  • Boundary markers: None; the instructions explicitly direct the agent to "Use DESIGN.md as the source of truth for UI decisions" and "Follow color roles, typography hierarchy, spacing scale, and component states from DESIGN.md."
  • Capability inventory: The agent possesses full capability to modify project source code and implement UI components based on the ingested content.
  • Sanitization: There is no sanitization, escaping, or validation performed on the downloaded markdown content before it is interpolated into the agent's context.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Aug 19, 2026, 04:38 PM
Security Audit — agent-trust-hub — awesome-design-md