# FeatureQL benchmarks catalog

Evaluation assets for FeatureQL. Not part of the documentation website sidebar.

## Publication boundary

- **Public specification:** `0_questions.md`, `1_data_setup.slt`, scoring notes, optional `2_hints.md`.
- **Reference:** `3_solutions.slt` / `3_solutions.md` (disclose only when grading allows).
- **Run results:** `5_answers_<model>.slt`, `6_postmortem_<model>.md` — per-model, not part of the definition.
- **Private holdouts:** paths with `.skip` / `.private` are excluded from this catalog.

Source tree: `docs/website/authored/benchmarks/`.

## Suites

| Suite | Slug | Path | Questions | Setup | Solutions | Eval package |
|---|---|---|---|---|---|---|

## Expected role files

- `benchmark.yaml`
- `0_questions.md`
- `1_data_setup.slt`
- `1_feature_setup.slt`
- `2_hints.md`
- `3_solutions.slt`
- `3_solutions_sql.slt`
- `3_solutions.md`
- `9_data_teardown.slt`