uiverify-triage-visual-changes

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.30). The skill reads review comments and AI verdicts from a UI Verify MCP server (https://uiverify.ai/mcp) via tools like list_comments and get_build, where comments can be authored by reviewers/outsiders.

MEDIUM W009: Direct money access capability detected (payment gateways, crypto, banking).

• Direct money access detected (medium risk: 0.30). The skill documents tools like review_diff and accept_build (SKILL.md:74-75, 158-161) that authorize visual regression baseline acceptances/decisions on a service, constituting a direct financial or operational execution (change of baseline state/approval authorization). However, the service access relies on a user OAuth login in the browser (SKILL.md:32, 289) and authenticated MCP server session credentials, establishing independent service-layer authentication and authorization checks. Therefore, it is a direct operation with qualifying independent access control.