Adds `eval.yaml` using the registered `extractbench` framework, with the overall mean and six challenge groups. Restores `rl-environment` alongside `harbor` to test Benchmark/RL badge coexistence.
Only `eval.yaml` and the dataset-card tag change; data and Harbor tasks are unchanged. YAML and task mappings were checked against the existing ExtractBench registration and Pro dataset configuration. Official Benchmark activation may require HF allowlisting; badge precedence remains to be checked after merge.
- Add Harbor tasks using the published ExtractBench coding-agent prompt (8b5daf42e2adb5e8ef6f13d0bc4366732cd7970f)
- Document Git LFS prerequisite for Harbor (092b835ca9918faf75c8c97e041326a10f2686ce)
- Add aligned cell feedback for the existing Value F1 reward (f32fd1d7d5056ee12286afa70e4d564fc353102d)
- Enforce agent-phase model-host allowlists and offline verification (f5eb7df6f3f9ceb635e6cec805e10ee28c68bb03)
Subsets, folders, row ids and challenge tags now follow the six challenges: document conventions, house vocabulary, house counting rules, drawn relations, charts inside documents and long chart sheets. Documents, schemas, expected outputs and field rules are unchanged.
