feat: add adaptive context productivity evaluation

[Coding-Dev-Tools]

Summary

• add adaptive prompt-context routing with safe bypass, retrieval, and fallback modes
• preserve session scope, temporal filters, content-free receipts, and accurate routing savings
• add a reproducible offline productivity benchmark with correction-loop, telemetry, and model-provenance safeguards

Validation

• python -m pytest tests/ -q --ignore=tests/test_hosted_evidence.py --ignore=tests/test_hosted_ledger.py --ignore=tests/test_hosted_luna.py
• python -m pytest tests/test_adaptive_context.py tests/test_productivity_eval.py tests/test_receipts.py -q
• ruff check engraphis/core/__init__.py engraphis/core/adaptive_context.py engraphis/core/engine.py engraphis/core/grounded.py engraphis/core/store.py engraphis/service.py eval/productivity.py tests/test_adaptive_context.py tests/test_productivity_eval.py
• python -m eval.productivity --dataset eval/datasets/codemem.jsonl --max-context-tokens 512 --retrieval-token-budget 256

Unrelated, untracked private-evaluation work remains intentionally excluded from this PR.