chore: scenario additions should be patch changesets, not minor

[sepehr-safari]

PR #133 shipped with a minor changeset, so the release PR #141 proposes
0.5.0.

v0.5.0 is reserved for OCPP 2.0.1. It is named as the next milestone in both
CURRENT_STATE.md and ROADMAP.md. Publishing it as "one new scenario" spends
that number on test corpus and pushes the 2.0.1 work to 0.6.0, breaking the
milestone-to-version mapping the project has held since v0.1.

The judgement call

A new scenario does add a public export, so minor is defensible read against
semver alone. But scenarios are test corpus rather than consumer API surface.
Nothing is built against an individual scenario constant; the registry is
consumed as a whole through scenarios and getScenario(). Treating scenario
additions as patch keeps the version numbers meaning what the roadmap says they
mean.

Fix

• Retag .changeset/warm-dragons-smile.md from minor to patch, so chore: version packages #141
  recomputes to 0.4.3.
• Write the convention into the "Adding a Scenario" section of CONTRIBUTING.md,
  which currently does not mention changesets at all. feat(scenarios): add firmware update success scenario #133 was asked for a minor
  in review, which is how this arrived, and Add a new scenario: heartbeat-timeout #137, Add a new scenario: firmware-update-failure #138 and Add a new scenario: repeated-boot-notification #139 are queued behind
  it.

Nothing has published yet: #141 is unmerged.