# Exact claims supported and not supported

## Supported by this artifact

- Thomas can design a readable mock evaluation and trace-review workflow for AI or agent QA.
- Thomas can structure UI states for evaluation runs, trace details, tool-call review, and human review boundaries.
- Thomas can use synthetic data labels and claim controls to keep a demo public-safe.
- Thomas can pair a UI proof with reviewer notes, approve or revise or blocked labels, and unsupported-claim removal guidance.
- Thomas can create source-backed launch-note and technical-review scaffolds for developer-facing content systems.
- Thomas can separate synthetic demo evidence from production product, customer, platform, or API claims.

## Not supported by this artifact

- Production Arize Phoenix experience.
- Production OpenRouter product usage, API key ownership, billing access, deployment, or customer adoption.
- Production AI observability, evaluation infrastructure, tracing infrastructure, or model-monitoring operation.
- Production React, TypeScript, accessibility, design-system, or frontend architecture depth beyond a small demo.
- Named customer work, case studies, revenue outcomes, or testimonials.
- 5 to 7 years of product marketing, viral technical content, large developer audience, founder ghostwriting, or enterprise PMM ownership.
- Work authorization, sponsorship, residency, tax, background check, KYC, e-signature, or binding employment authority.

## Safe positioning sentence

A safe low-claim sentence is: "I built a small synthetic React and TypeScript style proof that shows evaluation runs, trace review with mock tool calls, and a human review queue, paired with claim controls that prevent unsupported production or platform claims."

## Unsafe positioning to avoid

Do not say: "I have production Arize or OpenRouter observability experience," "I shipped this for customers," "this proves senior frontend depth," or "this is a deployed evaluation platform."