accessibility-audit
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill's setup process involves installing the
playwrightandaxe-corepackages via npm. These are reputable, industry-standard libraries used for browser automation and accessibility testing. - [COMMAND_EXECUTION]: The skill executes a local script,
../dogfood/scripts/browser-driver.mjs, using Node.js to perform browser automation tasks such as navigating to pages, capturing screenshots, and running accessibility scans. - [INDIRECT_PROMPT_INJECTION]: The skill processes content from external websites, creating a surface for indirect prompt injection where malicious instructions could be embedded in the audited pages.
- Ingestion points: The skill navigates to external URLs and captures accessibility snapshots and trees during Phase 2 (Automated Scan) and Phase 3 (Manual Verification).
- Boundary markers: The instructions do not provide explicit delimiters or warnings to isolate the agent from instructions that might be embedded in the target web content.
- Capability inventory: The skill possesses capabilities for browser automation, network navigation, and local file system writes (generating JSON findings, PNG screenshots, and Markdown reports).
- Sanitization: No sanitization or filtering of the website content is described before the agent analyzes it.
Audit Metadata