screen-reader-testing
Pass
Audited by Gen Agent Trust Hub on Feb 27, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill primarily consists of instructional content and checklists intended for accessibility auditing. It does not perform any automated background operations or modify system configurations.
- [EXTERNAL_DOWNLOADS]: The resource file contains links to official documentation and guides from trusted sources including Apple, NV Access, Freedom Scientific, and WebAIM. These are documented neutrally as legitimate references for the skill's primary purpose.
- [COMMAND_EXECUTION]: The documentation lists keyboard shortcuts and CLI commands specifically for controlling screen reader software (e.g., 'Insert + Q' for NVDA). These are presented as educational instructions for the operator and do not involve the skill executing arbitrary system commands.
- [PROMPT_INJECTION]: The skill's primary function is to evaluate web application code provided by users. This ingestion of untrusted data represents an indirect prompt injection surface; however, the skill provides specific structures (checklists and patterns) to guide the agent, and no active exploits were found in the provided files.
Audit Metadata