UI/UX Engineering Risk

Purpose

Protect the correctness and safety behind an interface without turning a UI/UX workflow into a full security audit. Activate this specialist only when the requested change crosses backend, data, API, trust, deployment, or high-impact automation boundaries.

Boundary for users: This skill provides design/review guidance, risk framing, negative-case planning, and evidence prompts. It is not automated security tooling and does not replace SAST, SCA, DAST, secret scanning, sandboxing, network-egress controls, MCP conformance testing, token revocation, runtime policy enforcement, continuous telemetry, penetration testing, or a qualified security/platform review. Teams must provide and run the controls required by their stack, deployment, threat model, and compliance obligations; record missing controls as residual risk instead of claiming the product is secure or production-ready.

Trigger decision

Activate engineering-risk when any of the following is true:

• The task changes or reviews database schema, migrations, queries, transactions, persistence, caching, or shared server state.
• The task changes an API, external service, upload/parser, server-side fetch, webhook, background job, or paid/limited resource.
• The task touches authentication, authorization, tenant/resource isolation, roles, permissions, secrets, personal data, or destructive actions.
• The task accepts untrusted input or output that may reach SQL, OS commands, templates, HTML, URLs, serialization, logs, or tools.
• The product embeds an LLM, retrieval, memory, agent, MCP/connector, or autonomous/high-impact action.
• The task includes deployment, dependencies, configuration, observability, production readiness, or a material security unknown.