Evaluate a zero-account browser playground after measuring the corrected beginner path

## User impact and evidence

A browser playground or small visual authoring surface could reduce discovery cost for people who cannot immediately run Docker. Comparative research identifies useful zero-account examples, but the corrected Registry Stack path has not yet been measured with independent readers.

## Why this is non-blocking now

The recovery must first prove the released CLI and Docker Compose path. Adding a browser executor or visual builder now would introduce another runtime, security boundary, and documentation rail without evidence that it is necessary.

## Owner

Registry Stack product and documentation maintainers, with security review.

## Prerequisites

- Completed fresh-reader and country-developer runs against the corrected path.
- A measured friction report that separates download time, container availability, authoring complexity, and documentation defects.
- A threat model for untrusted input, source values, secrets, persistence, sharing, and execution.

## Reconsideration trigger

Promote discovery when human evidence shows that the supported container contract prevents the target audience from reaching the first protected result, or when a country environment cannot use containers.

## Expected verification

- No account, personal data, country endpoint, or production credential is required.
- The surface cannot widen disclosure or turn detected fields into released fields.
- Shared examples are synthetic and cannot persist secrets or source values.
- Browser results match the canonical compiler and validators.
- The existing CLI journey remains supported and does not fork into a second configuration model.

No files, tests, or entry points are named. Start by completing the fresh-reader and country-developer runs against the corrected CLI and Docker Compose path, then use the friction report and threat model to determine whether a zero-account browser surface is warranted; done means the listed security, canonical-result, and CLI-compatibility conditions are verified.

Written by the indexing model from the issue text.