wcag-accessibility-audit
Pass
Audited by Gen Agent Trust Hub on Oct 10, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted data from external websites and screenshots, which serves as a vector for indirect prompt injection attacks.
- Ingestion points: The
urls_or_screenshotsinput field inSKILL.mdallows third-party web content to be introduced into the agent's processing context. - Boundary markers: The skill provides explicit instructions to the agent to use
<untrusted-content>delimiters to isolate external data and to treat the skill's own instructions as authoritative over any content found in the input. - Capability inventory: The audit procedure describes a workflow involving web fetching, rendering, and interaction with UI components via browser tools and assistive technology simulators.
- Sanitization: The 'Security Notice' section includes a heuristic checklist for the agent to identify and ignore common injection patterns, such as 'ignore previous instructions' or 'you are now', when they appear within the audited content.
Audit Metadata