Added invalid handling

[bryantaustin13]

Tests with "invalid": "semantic", handled as purposely failing as a valid test.

---

Moved from #98, which was raised from a fork before I had committer access on this repository. Same work, same branch name, now hosted here so CI and reviews run against cqframework directly.

[bryantaustin13]

Superseded by #121.

Re-evaluated this against current main and most of it has been overtaken:

• The parsedExpected !== undefined guard is already on main (run-test-core.ts), identical.
• The singletonListKeys guard is unnecessary — singletonListKeysFromExpected already returns an empty set for a non-array input.
• isExpectedInvalidResult() is superseded by Fix error-expecting test handling in CLI and server runners #116's responseIndicatesError, which checks the response status and looks structurally for an evaluation error parameter rather than regex-matching the stringified actual. Fix error-expecting test handling in CLI and server runners #116's version is the better one and is already merged.
• Letting invalid tests run without an <output> is already on main (results-shared.ts).

What remained was the capability normalization, which is a real and unfixed bug: a test declaring exactly one <capability> parses to a bare object rather than an array, so Result dropped it — 7,390 tests in the suite versus 513 with two or more. #121 is that fix on its own, with tests, on a branch name that describes it (normalize-single-capability). Since a PR's head branch cannot be repointed, it had to be a new PR.

Closing in favour of #121 — please continue any discussion there.