accesslint-diff
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content from live web pages and source code revisions which could contain malicious instructions designed to influence the agent.\n
- Ingestion points: The agent reads "live page" DOM data and revision content to identify violations (SKILL.md).\n
- Boundary markers: There are no requirements for using delimiters or boundary markers to isolate processed DOM data in reports.\n
- Capability inventory: The skill manages local servers, browser processes, and Git worktrees (SKILL.md).\n
- Sanitization: No instructions are provided for escaping or validating the "evidence" extracted from the page content.
Audit Metadata