design-code

Pass

Audited by Gen Agent Trust Hub on Aug 24, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to run multiple local verification scripts, such as scripts/lint_hardcodes.py, scripts/validate_theme_refs.py, scripts/contrast.py, node scripts/taste_audit.mjs, node scripts/verify_states.mjs, and node scripts/accuracy_report.mjs, to audit the generated code for correctness and design standards.- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied framework and styling specifications as inputs for its code generation and subsequent script-based verification workflow, creating a potential surface for indirect prompt injection.
  • Ingestion points: User-defined target framework and styling method from the initial request (SKILL.md).
  • Boundary markers: No explicit boundary markers or instructions to disregard embedded instructions within the user input are provided.
  • Capability inventory: Local shell access for running Python and Node.js verification tools (SKILL.md).
  • Sanitization: No sanitization or validation logic is specified for the user-provided framework and styling strings.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 24, 2026, 12:50 PM
Security Audit — agent-trust-hub — design-code