Context

browsertrace compare <failed_run_id> <success_run_id> is now available in v0.1.19. The README and Browser Use guide explain the command, but first-time users still need a tiny expected-output example so they know what a useful comparison result looks like.

Scope

Keep this docs-only and small. Suggested files:

• examples/README.md
• tests/test_metadata.py

Avoid generated release assets, broad README rewrites, UI changes, or launch copy changes.

Acceptance criteria

• Add a short browsertrace compare <failed_run_id> <success_run_id> expected-output example to examples/README.md.
• The example should show the first divergent step and at least one of these fields: action, url, status, or error.
• Mention that --json is available for automation, but do not paste a large JSON blob.
• Add or update a metadata test so this expected-output example stays documented.
• Do not add requests for stars, upvotes, reposts, or reciprocal promotion.

Verification

Run:

uv run --python 3.11 --extra dev pytest tests/test_metadata.py -q

A good first PR can follow the First PR Recipe:
https://github.com/aaronlab/browsertrace/blob/main/CONTRIBUTING.md#first-pr-recipe