test: refresh product os reader quality reference

[Stahl-G]

Summary

• extend the packaged same_evidence_reader_quality_regression case to generate clean delivery/audit bundle archives through a new eval-only packs.bundle action
• add source appendix and source appendix trace fixture artifacts plus finalize audit/delivery hash bindings
• add a public-safe v0.11.3 Product OS reader-quality reference note and link it from README/docs indexes

Boundary

• reference package / deterministic regression only
• no new runtime warning surface, projection artifact, gate authority, UI, semantic proof, output-quality score, delivery approval, or release authorization
• fixture remains synthetic public-safe data

Validation

• python3 -m pytest -q tests/test_evaluation_cases.py tests/test_report_bundles.py
• PYTHONPATH=src python3 -m multi_agent_brief.cli.main eval-cases validate --json
• PYTHONPATH=src python3 -m multi_agent_brief.cli.main eval-cases run --case-id same_evidence_reader_quality_regression --repo-workdir . --json
• python3 scripts/check_public_safety.py --path docs/reference-runs/v0.11.3-product-os-reader-quality-reference.md
• python3 scripts/check_public_safety.py --path src/multi_agent_brief/evaluation_cases/fixtures/cases/same_evidence_reader_quality_regression
• python3 scripts/check_product_baseline.py
• python3 scripts/check_release_consistency.py --no-tag
• python3 scripts/check_briefloop_skill_freshness.py
• python3 scripts/check_skill_contract.py
• python3 scripts/check_version_consistency.py
• PYTHONPATH=src python3 scripts/check_capabilities.py
• python3 scripts/check_runtime_asset_parity.py
• python3 scripts/generate_agent_configs.py --check
• python3 scripts/sync_hermes_plugin_skills.py --check
• git diff --check
• python3 -m compileall -q src tests
• non-editable install smoke: eval-cases validate + same_evidence_reader_quality_regression
• python3 -m pytest -q