pixel-audit
Pass
Audited by Gen Agent Trust Hub on Oct 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external sources including Figma nodes and rendered page markup to perform visual audits, creating an attack surface for instructions potentially embedded in audited UI content.
- Ingestion points: Figma MCP metadata, variables, and the DOM/CSS content of the target web page as specified in SKILL.md and references/evidence-capture.md.
- Boundary markers: No specific delimiters or warnings to ignore embedded instructions are provided for the interpolation of design or markup data.
- Capability inventory: The skill writes inventory and defect artifacts to the local file system, executes shell commands (curl, grep), and performs browser-based JavaScript evaluation.
- Sanitization: The instructions do not specify sanitization or filtering methods for content ingested from design tools or page sources.
- [COMMAND_EXECUTION]: The skill utilizes local shell commands to verify the asset pipeline and served content.
- Evidence: Use of curl to inspect built assets and grep to verify specific CSS classes or tokens within those assets (SKILL.md, references/evidence-capture.md).
- [DYNAMIC_EXECUTION]: The skill executes dynamic JavaScript within a browser context to extract element properties for verification.
- Evidence: Evaluation of getBoundingClientRect() and getComputedStyle() on DOM elements to prove visual fixes during the verification gate (SKILL.md, references/evidence-capture.md).
Audit Metadata