accessibility-review
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted external data in the form of code components, templates, and styles provided by the user for accessibility review.
- Ingestion points: The skill instructs the agent to review user-provided files such as components, templates, and test files (
SKILL.md). - Boundary markers: The skill specifies a structured output format (
## Findings) and includes principles to prevent immediate code modification, which helps separate the analysis from the execution context. - Capability inventory: The agent uses file-reading and analysis capabilities to perform the review. No active file-writing or network exfiltration tools are explicitly requested in the skill configuration.
- Sanitization: No explicit sanitization or filtering of the input code is defined, relying on the agent's internal safety guardrails.
Audit Metadata