Add admin defaults for new chat model and reasoning effort

[Chad Palmer (chadpalmer)]

Summary

Adds admin-configurable defaults for new non-agent chat conversations:

• Adds an opt-in Admin Settings toggle for applying a default model to new chats.
• Adds an admin default reasoning effort setting for new chats.
• Preserves the existing user “last selected model/reasoning effort” behavior while the admin override is off.
• Keeps existing conversations on their own last-used model instead of replacing them with the admin default.
• Handles the left-nav New Chat flow where /chats?conversationId=... points at a freshly created blank conversation.

Implementation Notes

• New app settings:
  • enable_default_model_for_new_conversations
  • default_reasoning_effort
• Uses the existing multi-endpoint default model selector for the admin default model.
• Restores existing conversation model state from saved message metadata.
• Treats blank conversations with no model history as new-chat default candidates.
• Prevents programmatic selector changes from being saved as the user’s preferred model.

Validation

Passed:

• git diff --check upstream/Development...HEAD
• python -m py_compile for changed Python app/test files
• node --check for changed chat JavaScript files
• python scripts/check_swagger_routes.py application/single_app/route_backend_chats.py application/single_app/route_frontend_admin_settings.py application/single_app/route_frontend_chats.py
• python scripts/check_xss_sinks.py --base-sha upstream/Development --head-sha HEAD ...
• python scripts/check_broken_access_control.py --base-sha upstream/Development --head-sha HEAD ...
• python functional_tests/test_default_model_selection_fallback.py
• python functional_tests/test_chat_model_selector_initial_multiendpoint_bootstrap.py
• python functional_tests/test_chat_template_json_bootstrap_safety.py
• python functional_tests/test_admin_settings_field_contract.py
• python functional_tests/test_docs_app_surface_coverage.py
• python functional_tests/test_docs_site_quality.py
• Route policy tests:
  • functional_tests/route_tests/test_route_blueprint_policy_inventory.py
  • functional_tests/route_tests/test_route_unauthenticated_policy_contract.py
  • functional_tests/route_tests/test_route_policy_test_coverage.py

Docs and Version

• Version bumped to 0.261.044.
• Admin docs updated.
• Release notes updated under v0.261.044.
• Docs app-surface inventory regenerated and verified.

Known Gaps

• Full authenticated browser UI tests were not run locally. The specific behavior was manually verified and covered with focused functional/source-contract tests.

[Chad Palmer (chadpalmer)]

Paul Lizer (@paullizer) Bionic711 This adds an admin default model and reasoning effort for new chats. This feature defaults to off so that the current behavior stays the same unless an admin turns the feature on and selects a default model and reasoning level for new chats. It looks like this completes the following roadmap task: #606