design-qa

Pass

Audited by Gen Agent Trust Hub on Jul 14, 2026

Risk Level: SAFE
Full Analysis
  • [Data Ingestion Surface]: The skill involves the ingestion of data from external sources, specifically through navigating to user-provided local or deployed URLs and interacting with Figma nodes to capture visual evidence.
  • Ingestion points: The workflow in SKILL.md (Step 2) directs the agent to open target implementations in a browser and capture screenshots.
  • Boundary markers: There are no explicit instructions or delimiters used to separate user-provided content from the agent's core instructions during the analysis process.
  • Capability inventory: The agent's capabilities include browser navigation, taking screenshots, and writing a markdown report to the project root (design-qa.md).
  • Sanitization: The instructions do not specify a requirement to sanitize or filter content found within the rendered HTML of target implementations before processing.
  • [File System Interaction]: The skill includes instructions to create and update a design-qa.md file to document the QA findings. This is a common and intended behavior for a documentation-oriented skill.
  • [External Reference]: The skill references a critical-overrides.md file via a relative path. This indicates the skill is designed to work within a specific project structure where shared references are available.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 14, 2026, 12:48 PM
Security Audit — agent-trust-hub — design-qa