omh-accessibility-audit
Pass
Audited by Gen Agent Trust Hub on Nov 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted external data, such as DOM snapshots, source code, and accessibility trees, to perform UI audits. This creates a surface where malicious instructions embedded in the analyzed data could attempt to influence the agent's verdict.
- Ingestion points: In
SKILL.md, the skill lists target application code, DOM snapshots, and accessibility trees as required inputs. - Boundary markers: The skill includes strong instructions to separate findings by Rule ID and maps them to concrete WCAG criteria. It explicitly forbids claiming a PASS status without fresh, observed evidence, which serves as a procedural boundary.
- Capability inventory: The skill is a reviewer role and does not request tools for file writing, network operations, or shell execution.
- Sanitization: There is no mention of explicit sanitization or escaping for the ingested UI data; the skill relies on structured reporting and human-readable evidence gates to mitigate influence.
Audit Metadata