design-to-code

Pass

Audited by Gen Agent Trust Hub on Sep 11, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests instructions and data from external design specifications and HTML files which could be manipulated to influence the agent's actions.
  • Ingestion points: docs/designs/.md, docs/designs/stack.md, and HTML content from git branches.
  • Capability inventory: Execution of git commands (fetch, show, push), bun scripts, and pixel-review tools.
  • Boundary markers: The skill lacks explicit delimiters or instructions to ignore embedded commands within the design documents.
  • Sanitization: No sanitization of the content from design files or HTML is described before the agent processes it.
  • [COMMAND_EXECUTION]: The skill executes shell commands that interpolate variables derived from file and branch names, creating a potential command injection surface.
  • Evidence: The variable is interpolated into git fetch, git show, and git push --delete commands.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 11, 2026, 10:58 PM
Security Audit — agent-trust-hub — design-to-code