design-system

Pass

Audited by Gen Agent Trust Hub on Aug 16, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is designed to ingest and process untrusted external data, which introduces a surface for indirect prompt injection attacks where malicious instructions could be embedded in input data to influence agent behavior.\n
  • Ingestion points: Untrusted data enters the agent context through project briefs in the 'Design read' phase and reference URLs passed to the 'extract-theme' skill in the 'Foundation branch' phase.\n
  • Boundary markers: The instructions lack explicit delimiters or warnings to the agent to disregard instructions potentially hidden within the user-provided briefs or site content.\n
  • Capability inventory: The skill has the ability to write to the file system (DESIGN.md and PAGES.md) and invoke other functional skills such as 'grilling' and 'prototype'.\n
  • Sanitization: No sanitization, filtering, or validation mechanisms are defined to process the external content before it is used to drive the design process.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 16, 2026, 10:48 PM
Security Audit — agent-trust-hub — design-system