consolidate scenarios

currently scenarios are written like unit tests. very small and isolated. however running every scenario takes ages with bazel. we should test multiple things per scenario. this will implicitly make the scenarios themselfes more realistic.

hint: attempt to follow realistic score scenarios

---

included or follow-up:

gap: add scenario with pytest and gtest. with a passing, failing, skipped, and errored testcase. Based on #774 the parser should be way more defensive, e.g. unknown attribute = error.

Start by locating the existing scenario tests and their Bazel target, then compare them with realistic score scenarios. Completion means scenarios exercise multiple behaviors while avoiding the current per-scenario runtime cost; pytest and gtest coverage, including passing, failing, skipped, and errored cases, is listed as included or follow-up work alongside the defensive parser changes from #774.

Written by the indexing model from the issue text.