Expand the eval corpus toward 150 pairs / more libraries

[AshwinUgale]

Why

The README roadmap targets a 150-pair x 5-library corpus. More pairs = tighter, more trustworthy accuracy/version/refusal metrics. This is approachable for anyone who can write good documentation Q/A - little core code required.

What to do

• Add well-formed entries to evals/corpus.json following the existing schema (evals/schema.py), including out-of-scope entries so the refusal metric stays honest.
• Cover a new library once its doc source lands (see Add a doc source for a new library: Flask #11 / Add a doc source for a new library: Django #12).

Acceptance criteria

• New entries validate against the schema: uv run pytest -c ../evals/pytest.ini ../evals/tests.
• Balanced in-scope / out-of-scope split noted in the PR.

Pointers

• evals/corpus.json, evals/schema.py, evals/README.md.