accessibility

Pass

Audited by Gen Agent Trust Hub on Sep 5, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill contains a data ingestion surface where it reads project-specific expectations from a local file.
  • Ingestion points: Reads accessibility expectations from .ux/CONTEXT.md when present.
  • Boundary markers: The skill does not define specific delimiters for the content read from this file, but it instructs the agent to ground work in available evidence and avoid inventing needs.
  • Capability inventory: The skill body contains only natural language instructions; there are no associated scripts, subprocess calls, or network operations.
  • Sanitization: No explicit sanitization or filtering of the file content is mentioned.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 5, 2026, 02:55 AM
Security Audit — agent-trust-hub — accessibility