compute-control and semantic loot closeout

[TacoRocket]

What changed

• tightened chain-family table wording and grouped output behavior to better match the family truthfulness baseline
• added additive CollectionIssue.scope support so issue location is first-class in the contract while preserving compatibility with existing collectors and goldens
• added regression coverage for scoped issues plus help, CLI, terminal UX, and visibility behavior around the updated wording

Why it changed

• Kube surfaced that AzureFox's operator-facing issue scope was stronger in table copy than in the JSON/model contract
• the family-style alignment pass also called for clearer current-scope wording, partial-visibility language, and more explicit grouped-output boundaries

Impact

• AzureFox output is more explicit about what is currently visible versus what is still unconfirmed
• downstream consumers can now rely on issue.scope instead of inferring issue location from prose alone
• the new regression tests give us a tripwire for future collectors that emit issues without scope

Validation

• python3 -m ruff check src tests scripts
• PYTHONPATH=src python3 scripts/generate_schemas.py
• PYTHONPATH=src python3 -m pytest -o cache_dir=/private/tmp/pytest-cache-dns tests/test_collectors.py tests/test_golden_outputs.py tests/test_contract_schemas.py tests/test_cli_smoke.py tests/test_help.py tests/test_terminal_ux.py tests/test_models.py
• PYTHONPATH=src python3 -m pytest -o cache_dir=/private/tmp/pytest-cache-all