databricks-agentbricks 0.9.0__tar.gz → 0.11.0__tar.gz

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (239) hide show
  1. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/PKG-INFO +131 -17
  2. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/README.md +130 -16
  3. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/pyproject.toml +1 -1
  4. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/cli/app.py +3 -3
  5. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/cli/deploy.py +118 -15
  6. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/cli/dev.py +6 -1
  7. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/cli/doctor.py +2 -3
  8. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/cli/memory.py +2 -0
  9. databricks_agentbricks-0.11.0/src/databricks_agentbricks/cli/memory_changes.py +257 -0
  10. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/cli/models.py +25 -0
  11. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/cli/permissions.py +8 -1
  12. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/cli/pipeline.py +16 -3
  13. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/cli/telemetry.py +177 -34
  14. databricks_agentbricks-0.11.0/src/databricks_agentbricks/clients/agent_details_client.py +86 -0
  15. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/clients/app_probe.py +6 -5
  16. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/clients/apps_client.py +49 -15
  17. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/clients/conversation_store_client.py +22 -0
  18. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/clients/databricks_cli.py +84 -15
  19. databricks_agentbricks-0.11.0/src/databricks_agentbricks/clients/output_tail.py +124 -0
  20. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/clients/terminal_relay.py +8 -1
  21. databricks_agentbricks-0.11.0/src/databricks_agentbricks/clients/workspace_identity.py +37 -0
  22. databricks_agentbricks-0.11.0/src/databricks_agentbricks/diagnostics.py +91 -0
  23. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/errors.py +44 -5
  24. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/model_upgrade.py +12 -3
  25. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/presentation/deploy.py +95 -47
  26. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/presentation/group.py +112 -4
  27. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/presentation/help.py +25 -0
  28. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/presentation/render.py +13 -0
  29. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/projects/agent_project.py +64 -13
  30. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/projects/app_manifest.py +48 -23
  31. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/services/deploy_service.py +94 -51
  32. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/services/deployment/deploy_state.py +40 -7
  33. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/services/deployment/provisioners.py +35 -0
  34. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/services/dev_service.py +83 -71
  35. databricks_agentbricks-0.11.0/src/databricks_agentbricks/telemetry/__init__.py +1 -0
  36. databricks_agentbricks-0.11.0/src/databricks_agentbricks/telemetry/scrub.py +104 -0
  37. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/agent-bricks-workflow/SKILL.md +5 -2
  38. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/agent-bricks-workflow/references/deployment.md +18 -1
  39. databricks_agentbricks-0.11.0/src/databricks_agentbricks/templates/agent-bricks-workflow/references/model-upgrades.md +47 -0
  40. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/agent-bricks-workflow/references/state-and-tools.md +17 -0
  41. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/agent-langgraph/pyproject.toml +1 -1
  42. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/agent-openai/pyproject.toml +1 -1
  43. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/custom-agent-langgraph/agent/agent.py +2 -1
  44. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/custom-agent-langgraph/pyproject.toml +1 -1
  45. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/custom-agent-openai/pyproject.toml +1 -1
  46. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/_api_client.py +67 -1
  47. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/langgraph/messages.py +19 -1
  48. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/model_upgrades/optimization.py +244 -172
  49. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/.gitignore +0 -0
  50. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/NOTICE +0 -0
  51. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/__init__.py +0 -0
  52. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/agent_hint.py +0 -0
  53. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/cli/__init__.py +0 -0
  54. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/cli/__main__.py +0 -0
  55. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/cli/auth.py +0 -0
  56. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/cli/endpoint.py +0 -0
  57. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/cli/experimental.py +0 -0
  58. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/cli/init.py +0 -0
  59. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/cli/mcp.py +0 -0
  60. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/cli/prompts.py +0 -0
  61. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/cli/sandbox.py +0 -0
  62. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/cli/sessions.py +0 -0
  63. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/cli/tools.py +0 -0
  64. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/cli/tracing.py +0 -0
  65. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/clients/__init__.py +0 -0
  66. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/clients/api_client_provider.py +0 -0
  67. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/clients/apps_user_auth_client.py +0 -0
  68. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/clients/legacy_runtime_store.py +0 -0
  69. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/clients/local_ports.py +0 -0
  70. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/clients/local_tracing_client.py +0 -0
  71. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/clients/managed_runtime_store.py +0 -0
  72. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/clients/mlflow_trace_client.py +0 -0
  73. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/clients/tracing_client.py +0 -0
  74. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/presentation/__init__.py +0 -0
  75. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/presentation/dev.py +0 -0
  76. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/presentation/endpoint.py +0 -0
  77. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/presentation/loadtest.py +0 -0
  78. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/presentation/permissions.py +0 -0
  79. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/presentation/reporter.py +0 -0
  80. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/presentation/theme.py +0 -0
  81. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/presentation/tracing.py +0 -0
  82. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/projects/__init__.py +0 -0
  83. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/projects/config.py +0 -0
  84. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/projects/resolver.py +0 -0
  85. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/projects/types.py +0 -0
  86. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/py.typed +0 -0
  87. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/reporting.py +0 -0
  88. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/services/__init__.py +0 -0
  89. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/services/deployment/__init__.py +0 -0
  90. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/services/deployment/app_provisioner.py +0 -0
  91. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/services/deployment/config.py +0 -0
  92. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/services/deployment/model_binding_provisioner.py +0 -0
  93. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/services/deployment/names.py +0 -0
  94. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/services/deployment/tool_access.py +0 -0
  95. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/services/deployment/tool_access_provisioner.py +0 -0
  96. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/services/endpoint_service.py +0 -0
  97. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/services/invoke/__init__.py +0 -0
  98. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/services/invoke/auth.py +0 -0
  99. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/services/invoke/request.py +0 -0
  100. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/services/invoke/transport.py +0 -0
  101. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/services/loadtest/__init__.py +0 -0
  102. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/services/loadtest/_locustfile.py +0 -0
  103. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/services/loadtest/plan.py +0 -0
  104. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/services/loadtest/request.py +0 -0
  105. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/services/loadtest/runner.py +0 -0
  106. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/services/loadtest/streaming.py +0 -0
  107. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/services/tracing/__init__.py +0 -0
  108. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/services/tracing/targets.py +0 -0
  109. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/services/tracing_service.py +0 -0
  110. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/skills.py +0 -0
  111. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/agent-bricks-migrate/SKILL.md +0 -0
  112. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/agent-langgraph/.env.example +0 -0
  113. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/agent-langgraph/.gitignore +0 -0
  114. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/agent-langgraph/AGENTKIT_CONTRACT.md +0 -0
  115. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/agent-langgraph/AGENTS.md +0 -0
  116. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/agent-langgraph/CLAUDE.md +0 -0
  117. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/agent-langgraph/README.md +0 -0
  118. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/agent-langgraph/agent/__init__.py +0 -0
  119. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/agent-langgraph/agent/agent.py +0 -0
  120. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/agent-langgraph/agent/mcps.py +0 -0
  121. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/agent-langgraph/agent/tools/__init__.py +0 -0
  122. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/agent-langgraph/agent/tools/sample_tool.py +0 -0
  123. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/agent-langgraph/agent/tools/send_message.py +0 -0
  124. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/agent-langgraph/app.yaml +0 -0
  125. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/agent-langgraph/runtime/__init__.py +0 -0
  126. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/agent-langgraph/runtime/adapter.py +0 -0
  127. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/agent-langgraph/runtime/main.py +0 -0
  128. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/agent-langgraph/tests/__init__.py +0 -0
  129. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/agent-langgraph/tests/test_agent.py +0 -0
  130. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/agent-langgraph/tests/test_runtime.py +0 -0
  131. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/agent-langgraph/tests/test_workspace.py +0 -0
  132. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/agent-openai/.env.example +0 -0
  133. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/agent-openai/.gitignore +0 -0
  134. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/agent-openai/AGENTKIT_CONTRACT.md +0 -0
  135. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/agent-openai/AGENTS.md +0 -0
  136. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/agent-openai/CLAUDE.md +0 -0
  137. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/agent-openai/README.md +0 -0
  138. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/agent-openai/agent/__init__.py +0 -0
  139. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/agent-openai/agent/agent.py +0 -0
  140. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/agent-openai/agent/checkpoints.py +0 -0
  141. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/agent-openai/agent/mcps.py +0 -0
  142. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/agent-openai/agent/tools/__init__.py +0 -0
  143. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/agent-openai/agent/tools/sample_tool.py +0 -0
  144. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/agent-openai/agent/tools/send_message.py +0 -0
  145. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/agent-openai/app.yaml +0 -0
  146. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/agent-openai/runtime/__init__.py +0 -0
  147. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/agent-openai/runtime/adapter.py +0 -0
  148. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/agent-openai/runtime/main.py +0 -0
  149. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/agent-openai/tests/__init__.py +0 -0
  150. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/agent-openai/tests/test_agent.py +0 -0
  151. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/agent-openai/tests/test_checkpoints.py +0 -0
  152. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/agent-openai/tests/test_runtime.py +0 -0
  153. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/agent-openai/tests/test_workspace.py +0 -0
  154. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/custom-agent-langgraph/.env.example +0 -0
  155. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/custom-agent-langgraph/.gitignore +0 -0
  156. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/custom-agent-langgraph/README.md +0 -0
  157. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/custom-agent-langgraph/agent/__init__.py +0 -0
  158. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/custom-agent-langgraph/app.yaml +0 -0
  159. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/custom-agent-langgraph/runtime/__init__.py +0 -0
  160. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/custom-agent-langgraph/runtime/main.py +0 -0
  161. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/custom-agent-langgraph/tests/test_runtime.py +0 -0
  162. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/custom-agent-openai/.env.example +0 -0
  163. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/custom-agent-openai/.gitignore +0 -0
  164. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/custom-agent-openai/README.md +0 -0
  165. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/custom-agent-openai/agent/__init__.py +0 -0
  166. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/custom-agent-openai/agent/agent.py +0 -0
  167. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/custom-agent-openai/app.yaml +0 -0
  168. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/custom-agent-openai/runtime/__init__.py +0 -0
  169. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/custom-agent-openai/runtime/main.py +0 -0
  170. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/custom-agent-openai/tests/test_runtime.py +0 -0
  171. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/sandbox_mcp.py +0 -0
  172. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/sandbox_mcp_langgraph.py +0 -0
  173. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/ui/agent-langgraph/CHAT_APP.md +0 -0
  174. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/ui/agent-langgraph/runtime/main.py +0 -0
  175. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/ui/agent-langgraph/runtime/ui.py +0 -0
  176. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/ui/agent-langgraph/tests/test_demo_ui.py +0 -0
  177. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/ui/agent-langgraph/ui/agentbricks-mark.png +0 -0
  178. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/ui/agent-langgraph/ui/app.js +0 -0
  179. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/ui/agent-langgraph/ui/index.html +0 -0
  180. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/ui/agent-langgraph/ui/markdown-it-15.0.2.min.js +0 -0
  181. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/ui/agent-langgraph/ui/styles.css +0 -0
  182. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/ui/agent-openai/CHAT_APP.md +0 -0
  183. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/ui/agent-openai/runtime/main.py +0 -0
  184. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/ui/agent-openai/runtime/ui.py +0 -0
  185. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/ui/agent-openai/tests/test_demo_ui.py +0 -0
  186. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/ui/agent-openai/ui/agentbricks-mark.png +0 -0
  187. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/ui/agent-openai/ui/app.js +0 -0
  188. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/ui/agent-openai/ui/index.html +0 -0
  189. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/ui/agent-openai/ui/markdown-it-15.0.2.min.js +0 -0
  190. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentbricks/templates/ui/agent-openai/ui/styles.css +0 -0
  191. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/__init__.py +0 -0
  192. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/_pagination.py +0 -0
  193. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/_permissions.py +0 -0
  194. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/client.py +0 -0
  195. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/langgraph/__init__.py +0 -0
  196. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/langgraph/genie.py +0 -0
  197. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/langgraph/history.py +0 -0
  198. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/langgraph/mcp.py +0 -0
  199. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/langgraph/memory.py +0 -0
  200. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/langgraph/responses.py +0 -0
  201. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/langgraph/session_store.py +0 -0
  202. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/memory_store.py +0 -0
  203. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/model_upgrades/__init__.py +0 -0
  204. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/models.py +0 -0
  205. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/openai/__init__.py +0 -0
  206. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/openai/genie.py +0 -0
  207. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/openai/history.py +0 -0
  208. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/openai/mcp.py +0 -0
  209. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/openai/memory.py +0 -0
  210. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/openai/models.py +0 -0
  211. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/openai/sessions.py +0 -0
  212. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/py.typed +0 -0
  213. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/runtime/ARCHITECTURE.md +0 -0
  214. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/runtime/README.md +0 -0
  215. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/runtime/__init__.py +0 -0
  216. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/runtime/app.py +0 -0
  217. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/runtime/auth.py +0 -0
  218. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/runtime/cors.py +0 -0
  219. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/runtime/durability/__init__.py +0 -0
  220. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/runtime/durability/execution.py +0 -0
  221. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/runtime/durability/heartbeat.py +0 -0
  222. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/runtime/durability/lakebase_runtime_store.py +0 -0
  223. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/runtime/durability/recovery.py +0 -0
  224. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/runtime/durability/store.py +0 -0
  225. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/runtime/errors.py +0 -0
  226. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/runtime/execution.py +0 -0
  227. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/runtime/genie.py +0 -0
  228. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/runtime/mcp_auth.py +0 -0
  229. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/runtime/model_services.py +0 -0
  230. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/runtime/prompt_registry.py +0 -0
  231. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/runtime/runtime.py +0 -0
  232. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/runtime/session_store_client.py +0 -0
  233. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/runtime/store.py +0 -0
  234. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/runtime/tool_manifest.py +0 -0
  235. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/runtime/tracing.py +0 -0
  236. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/runtime/types.py +0 -0
  237. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/runtime/workspace.py +0 -0
  238. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/session_store.py +0 -0
  239. {databricks_agentbricks-0.9.0 → databricks_agentbricks-0.11.0}/src/databricks_agentkit/timefmt.py +0 -0
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.5
2
2
  Name: databricks-agentbricks
3
- Version: 0.9.0
3
+ Version: 0.11.0
4
4
  Summary: Agent Bricks CLI and AgentKit SDK for custom agents on Databricks
5
5
  Author-email: Databricks <agent-feedback@databricks.com>
6
6
  License-File: NOTICE
@@ -164,11 +164,32 @@ credentials.
164
164
 
165
165
  If Databricks SDK default authentication is already configured, you can skip `agentbricks login`.
166
166
  You can also pass the global `--profile/-p` option before an individual command, for example
167
- `agentbricks --profile <profile> tools list`. Use `--output json` for scripting.
167
+ `agentbricks --profile <profile> tools list`. Use `--output json` (before or after the command) for
168
+ scripting.
168
169
 
169
170
  The CLI sends best effort command usage telemetry after leaf commands. Its log records the static
170
171
  command path, package version, operating system, duration, and process outcome (exit code and, for
171
- nonzero exits, a bounded error category). Parameter entries include explicitly supplied options
172
+ nonzero exits, a bounded error category). Failures can also include up to 4096 UTF-8 bytes of
173
+ error text in `failure_message`. Error text includes exception messages and fully captured
174
+ failed subprocess output, with recognized credentials, paths, and email addresses scrubbed.
175
+ Combined failure text over 65,536 characters is also omitted before scrubbing.
176
+ Credential assignments, auth headers, credential options, and encoded text cause the remaining
177
+ captured text to be omitted; preceding diagnostic text is retained. This avoids parsing arbitrary
178
+ JSON, YAML, or Python values. Recognizable standalone tokens and private keys are also masked.
179
+ This follows the Asset Bundles minimization pattern and is best effort; other customer content
180
+ may remain in service responses or application output.
181
+ It is labeled `DATA_LABEL_USER_COMMANDS_RESPONSE`, excluded from centralized logfood, and retained
182
+ regionally only where customer logging and encryption policies permit. Error text is omitted for
183
+ parser errors and interruptions. Collection requires Legal approval and schema deployment.
184
+ For `dev`, failure text can include complete child output up to 3072 bytes per captured stream.
185
+ Once a stream exceeds that limit, all its retained text is permanently discarded and replaced with
186
+ `[REDACTED]`, including later plain error messages, because partial output may have lost credential
187
+ context. Short complete output is scrubbed before upload, including private keys. Displayed child
188
+ output is unchanged. With one output stream redirected, the terminal stream is inherited unchanged
189
+ and only the redirected stream is retained. When both streams point to the same regular file, both
190
+ are inherited without capture to preserve their original offsets and redirection behavior.
191
+ Startup milestones remain internal; a killed process or unavailable upload can still leave no event.
192
+ Parameter entries include explicitly supplied options
172
193
  and parsed Click defaults with supported typed values, automatically across all commands and
173
194
  parent groups. Declare defaults in Click options so the shared emitter can read the resolved
174
195
  values; fallbacks applied only inside command callbacks are not visible to Click. Omitted free-form
@@ -178,9 +199,8 @@ choices contribute an `AGENTBRICKS_CLI_`-prefixed value. Click boolean and numer
178
199
  contribute their typed values; explicitly supplied non-finite floats and integers outside the log
179
200
  schema's 64-bit range remain presence-only, while unsupported defaults are omitted. Hidden and
180
201
  deprecated options follow the same policy, and repeated
181
- options contribute one parameter entry per supplied value, up to 64 entries. Positional arguments,
182
- raw command arguments, paths, IDs, secrets, arbitrary free-form option values, and exception
183
- messages are omitted. Set
202
+ options contribute one parameter entry per supplied value, up to 64 entries. Parameter telemetry
203
+ omits positional arguments, raw command arguments, paths, IDs, and arbitrary free-form values. Set
184
204
  `AGENTBRICKS_DISABLE_TELEMETRY=1` to disable these uploads.
185
205
 
186
206
  Telemetry runs only when a leaf command is invoked, whether its callback succeeds or fails. Help
@@ -193,6 +213,25 @@ thread, returning sooner if the worker finishes. This is not a delivery guarante
193
213
  uploads can be lost when the process exits after the allowance expires. Missing credentials or a
194
214
  failed upload never change the command result.
195
215
 
216
+ Deploy and `deployments wait` telemetry includes a random invocation UUID, whether a gated
217
+ agent-details link was printed, the observed deployment state, and whether the process is an
218
+ internal background child. Details links carry the invocation UUID so browser arrivals can be
219
+ joined to the command that printed them. The UUID contains no agent name or authentication data.
220
+ Rollout outcomes are recorded independently of link availability, including failed rollouts;
221
+ preparation failures leave the deployment outcome unspecified.
222
+ `AGENTBRICKS_DISABLE_TELEMETRY=1` also removes these attribution parameters while keeping the link.
223
+ `deployments get` and `deployments list` also attach a per-command UUID to details links.
224
+ All rows from one list share one UUID and one command event; their deployment outcome stays
225
+ unspecified because these commands only read state. List links count as emitted only when JSON
226
+ output or terminal hyperlink rendering preserves their targets. Plain redirected output and
227
+ terminal modes that strip hyperlink targets keep `details_link_emitted=false`; use `--output json`
228
+ to capture the URLs. Group arrivals by the joined event's
229
+ `command_path` to distinguish deploy, wait, get, and list. Count distinct invocation IDs, rather
230
+ than list rows or repeated page loads, for command conversion percentages.
231
+ A wait command has its own invocation ID; it is not another deploy. Exclude background children
232
+ from user-invoked command counts. Link arrivals measure browser opens, including copied links,
233
+ not terminal mouse clicks or proof that the deployer and viewer are the same person.
234
+
196
235
  ## Quickstart
197
236
 
198
237
  The shortest path from a blank directory to a running and deployed agent:
@@ -228,6 +267,18 @@ The app's compute starts first and boots while the other resources are provision
228
267
  deploy stops partway, the app it created is kept: re-run `agentbricks deploy` to resume, or remove
229
268
  it with `agentbricks deployments delete agent-bricks-my-agent`.
230
269
 
270
+ When the workspace has both the Agent Memory Sessions beta and the new Agents UI enabled,
271
+ deploy output also includes an **Agent details** link to the workspace management page. This
272
+ applies to first deploys, redeploys, `--no-wait`, and `deployments wait`.
273
+ `deployments get` shows the same link alongside the app URL; `deployments list` links agent
274
+ names to their details pages and keeps a separate **Open app** link. JSON output adds
275
+ `details_url` when available and retains `url` for the deployed agent. The optional flag lookup
276
+ waits at most two seconds per command, including lists; unavailable or disabled flags omit the
277
+ link without affecting the command.
278
+ When following a local deploy record, `deployments wait` also requires its recorded workspace
279
+ host and ID to match the selected profile. Older records or records without a workspace ID omit
280
+ the management link; the recorded deployment status and invocation URL are still available.
281
+
231
282
  ### Deploying from a coding agent
232
283
 
233
284
  A first deploy takes ~3 min (mostly starting app compute); redeploys are faster. Either run
@@ -715,8 +766,8 @@ agentbricks memory stores permissions revoke memory-stores/<id> \
715
766
 
716
767
  Session stores accept their configured name or `session-stores/<name>`. Memory stores accept an ID
717
768
  or `memory-stores/<id>`; permission Get also accepts the display-name lookup supported by store Get
718
- and rejects ambiguous matches. The global `--profile` and `--output json` options go before
719
- `sessions` or `memory`. List follows the existing CLI pagination convention: one JSON/piped page
769
+ and rejects ambiguous matches. The global `--profile` option goes before `sessions` or `memory`;
770
+ `--output json` goes before or after the command. List follows the existing CLI pagination convention: one JSON/piped page
720
771
  with `next_page_token`, interactive text pages, or every page with `--all`.
721
772
 
722
773
  Get always returns the authenticated caller's effective permission, including
@@ -839,7 +890,10 @@ Use the permission commands below as the store creator or a workspace admin to r
839
890
  ## Model and prompt upgrades (experimental)
840
891
 
841
892
  `agentbricks experimental models` and `agentbricks experimental prompts` find the best model and
842
- prompt for each LLM call in your agent and swap them in. They live under `experimental` because
893
+ prompt for each LLM call in your agent and swap them in. Use them when a new model launches, or when
894
+ you want the same quality for less: you supply an eval set and how much quality, latency, and cost
895
+ matter, and the search tries each candidate model per LLM call, rewrites the prompts for it, and
896
+ recommends the combination that scores best on your whole agent. They live under `experimental` because
843
897
  their flags and output may still change between releases without a deprecation period. The
844
898
  agent.toml tables they write live under `[experimental]` too.
845
899
 
@@ -860,17 +914,56 @@ agentbricks experimental models status # the job's state, then its recomm
860
914
  agentbricks experimental models apply # switch to it
861
915
  ```
862
916
 
917
+ The `agent.eval:...` references name things in your project. A minimal `agent/eval.py`:
918
+
919
+ ```python
920
+ import mlflow
921
+
922
+ from agent.agent import answer # your agent's own entry point, called synchronously
923
+
924
+ mlflow.langchain.autolog() # traces token usage, which cost scoring needs; use your framework's autolog
925
+
926
+
927
+ def predict(inputs):
928
+ return answer(inputs["question"]) # wrap an async agent in asyncio.run(...)
929
+
930
+
931
+ TRAIN = [
932
+ {
933
+ "inputs": {"question": "Plan 3 days in Tokyo in late March"},
934
+ "expectations": {"expected_response": "..."},
935
+ },
936
+ # ... a few dozen records that look like real traffic
937
+ ]
938
+ VAL = [...] # same shape, held out; the winner and the summary's medians come from these
939
+
940
+
941
+ def mentions_expected(inputs, expectations, answer):
942
+ return float(expectations["expected_response"].lower() in str(answer).lower())
943
+
944
+
945
+ SCORERS = [mentions_expected] # or MLflow scorers, e.g. mlflow.genai.scorers.Correctness()
946
+ ```
947
+
948
+ Each of the agent's LLM calls must call its role's model service, `resolve_model_service("<role>")`
949
+ from `databricks_agentkit.runtime.tool_manifest`, so the search can route it to each candidate.
950
+
863
951
  By default `models upgrade` uploads the project to your workspace and runs the search as a
864
952
  serverless Databricks job, because a search can take hours. `--run-on local` runs it in your
865
953
  terminal instead, which needs the `upgrade` extra and the agent's dependencies installed. Each call
866
954
  takes longer from a laptop than on a serverless job, and a record slower than `--latency-gate` (60
867
955
  seconds by default) scores 0, so raise it if a local run warns that records went over the gate. The job imports the `predict_fn`,
868
956
  eval data, and scorers you name from the project (the same inputs `optimize_prompts_and_models`
869
- takes, below) and runs each eval record once per candidate, against a temporary `<service>_exp_<id>`
870
- clone. Each LLM call must use its role's service, `resolve_model_service("<role>")`, which reads the
957
+ takes, below) and runs each eval record once per candidate, against temporary clones of the model
958
+ services: one per service and model it may use, `<service>_exp_<id>_<model>`, created up front and
959
+ deleted when the search ends. Each clone always serves its one model, so switching candidates
960
+ redirects calls to another clone instead of waiting for a service's model to change. Each LLM call must use its role's service, `resolve_model_service("<role>")`, which reads the
871
961
  `AGENT_MODEL_SERVICE_<ROLE>` env var that deploy (and the job, before importing your code) sets.
872
- GEPA also rewrites every prompt bound with `prompts bind` for each candidate model. Pass
873
- `--prompt` to pick specific prompts instead, or `--models-only` to leave prompts alone. `models apply` switches
962
+ GEPA also rewrites every prompt bound with `prompts bind` for each candidate model, so you don't
963
+ pass prompt URIs unless you want a different set. Pass `--prompt` to pick specific prompts instead,
964
+ or `--models-only` to leave prompts alone. A `--prompt` URI must name an alias
965
+ (`prompts:/main.my_agent.writer@production`), not a version (`.../writer/3`): apply moves the
966
+ alias to the rewritten version, so an agent that loads a fixed version would never see it. `models apply` switches
874
967
  to the best quality / latency / cost trade-off and registers any rewritten prompts (prior versions
875
968
  keep `@production_previous`). `models rollback` undoes the whole last apply: it switches every
876
969
  model back and moves each prompt's alias back to its prior version. Apply refuses recommendations
@@ -879,6 +972,11 @@ recorded when the recommendation was applied. If rollback fails partway through,
879
972
  to finish the remaining changes. Each search owns its temporary clones, so overlapping jobs can
880
973
  evaluate different models independently.
881
974
 
975
+ When the search finishes, the summary shows each role's current and recommended model, the change in
976
+ the combined score, and how the median quality, latency, and cost per request moved from the
977
+ current setup to the recommendation on the validation records, for example
978
+ `Median cost per request $0.0040 → $0.0010 (-75%)`. `models status` shows the same summary later.
979
+
882
980
  The search is `databricks_agentkit.model_upgrades`, which you can also call directly (for example
883
981
  from a notebook) to tune several model services and MLflow Prompt Registry prompts together:
884
982
 
@@ -1128,7 +1226,7 @@ requests. By default, each App user's `X-Routing-Key` matches its generated user
1128
1226
 
1129
1227
  The final summary includes counts, throughput including drain, successful-only and all-completed
1130
1228
  latency percentiles, answer-text TTFT, token usage, coverage, incomplete attempts, and generator CPU
1131
- warnings. With `--output json` before `endpoint`, stdout contains one Agent Bricks summary object;
1229
+ warnings. With `--output json`, stdout contains one Agent Bricks summary object;
1132
1230
  progress stays on stderr. This is the same object saved as `summary.json`.
1133
1231
  Locust's rounded histogram buckets are used for percentiles; all-completed timings include failures.
1134
1232
  These bounded tests measure the selected workload and do not establish sustainable capacity or answer quality.
@@ -1514,9 +1612,8 @@ agentbricks tools remove genie_agent
1514
1612
  to `user`; choose `--auth app` deliberately for App service-principal execution. Existing manifests
1515
1613
  without `auth` preserve App/default identity. Both add commands and `remove` accept `--source PATH`
1516
1614
  to select a project instead of the current directory. Discovery needs no project; read that
1517
- project's `agent.toml` to inspect configured bindings. For scripted output, put the global
1518
- `-o json` option before `tools`, as in
1519
- `agentbricks -o json tools add genie-one --source ./my-agent`. Adding a binding is offline: it updates
1615
+ project's `agent.toml` to inspect configured bindings. For scripted output, add `-o json`, as in
1616
+ `agentbricks tools add genie-one --source ./my-agent -o json`. Adding a binding is offline: it updates
1520
1617
  `agent.toml` without contacting Genie or checking permissions. The corresponding sources are:
1521
1618
 
1522
1619
  ```toml
@@ -1608,3 +1705,20 @@ sync/streaming/background transport selector is manual.
1608
1705
  Developing Agent Bricks CLI (`agentbricks`), AgentKit, the runtime, and templates - plus the local dev loop and how to
1609
1706
  test unreleased changes on `agentbricks dev` and `agentbricks deploy`, is covered in
1610
1707
  [CONTRIBUTING.md](CONTRIBUTING.md).
1708
+
1709
+ The deployment workflow tracks phases and readiness internally; these observations are not
1710
+ logged as separate telemetry fields. Network errors are retried within the existing readiness budget; if the final probe remains unreachable,
1711
+ the deployed app has unverified readiness rather than a reported application failure. These
1712
+ observations do not validate model or tool behavior.
1713
+
1714
+ Command outcomes can be interpreted from the existing command path, exit code, error category,
1715
+ and options. Pending waits (`deployments wait` exit 75) and background submissions
1716
+ (`deploy --no-wait`) should not be counted as deployment failures. Deployment operation UUIDs
1717
+ stay in local state and the background worker; they are not uploaded. The separate uploaded
1718
+ `invocation_id` identifies one command and joins details-link arrivals to it.
1719
+ See [outcome interpretation](CONTRIBUTING.md#interpreting-command-and-deployment-outcomes) for
1720
+ failure-rate denominators and coverage limits.
1721
+
1722
+ Local runs track parent interruption and child exit status internally. A child-only SIGINT or
1723
+ explicit exit 130 remains a runtime failure; it is not treated as a confirmed user cancellation.
1724
+ Caught parent interruptions use the existing interruption category and omit failure text.
@@ -121,11 +121,32 @@ credentials.
121
121
 
122
122
  If Databricks SDK default authentication is already configured, you can skip `agentbricks login`.
123
123
  You can also pass the global `--profile/-p` option before an individual command, for example
124
- `agentbricks --profile <profile> tools list`. Use `--output json` for scripting.
124
+ `agentbricks --profile <profile> tools list`. Use `--output json` (before or after the command) for
125
+ scripting.
125
126
 
126
127
  The CLI sends best effort command usage telemetry after leaf commands. Its log records the static
127
128
  command path, package version, operating system, duration, and process outcome (exit code and, for
128
- nonzero exits, a bounded error category). Parameter entries include explicitly supplied options
129
+ nonzero exits, a bounded error category). Failures can also include up to 4096 UTF-8 bytes of
130
+ error text in `failure_message`. Error text includes exception messages and fully captured
131
+ failed subprocess output, with recognized credentials, paths, and email addresses scrubbed.
132
+ Combined failure text over 65,536 characters is also omitted before scrubbing.
133
+ Credential assignments, auth headers, credential options, and encoded text cause the remaining
134
+ captured text to be omitted; preceding diagnostic text is retained. This avoids parsing arbitrary
135
+ JSON, YAML, or Python values. Recognizable standalone tokens and private keys are also masked.
136
+ This follows the Asset Bundles minimization pattern and is best effort; other customer content
137
+ may remain in service responses or application output.
138
+ It is labeled `DATA_LABEL_USER_COMMANDS_RESPONSE`, excluded from centralized logfood, and retained
139
+ regionally only where customer logging and encryption policies permit. Error text is omitted for
140
+ parser errors and interruptions. Collection requires Legal approval and schema deployment.
141
+ For `dev`, failure text can include complete child output up to 3072 bytes per captured stream.
142
+ Once a stream exceeds that limit, all its retained text is permanently discarded and replaced with
143
+ `[REDACTED]`, including later plain error messages, because partial output may have lost credential
144
+ context. Short complete output is scrubbed before upload, including private keys. Displayed child
145
+ output is unchanged. With one output stream redirected, the terminal stream is inherited unchanged
146
+ and only the redirected stream is retained. When both streams point to the same regular file, both
147
+ are inherited without capture to preserve their original offsets and redirection behavior.
148
+ Startup milestones remain internal; a killed process or unavailable upload can still leave no event.
149
+ Parameter entries include explicitly supplied options
129
150
  and parsed Click defaults with supported typed values, automatically across all commands and
130
151
  parent groups. Declare defaults in Click options so the shared emitter can read the resolved
131
152
  values; fallbacks applied only inside command callbacks are not visible to Click. Omitted free-form
@@ -135,9 +156,8 @@ choices contribute an `AGENTBRICKS_CLI_`-prefixed value. Click boolean and numer
135
156
  contribute their typed values; explicitly supplied non-finite floats and integers outside the log
136
157
  schema's 64-bit range remain presence-only, while unsupported defaults are omitted. Hidden and
137
158
  deprecated options follow the same policy, and repeated
138
- options contribute one parameter entry per supplied value, up to 64 entries. Positional arguments,
139
- raw command arguments, paths, IDs, secrets, arbitrary free-form option values, and exception
140
- messages are omitted. Set
159
+ options contribute one parameter entry per supplied value, up to 64 entries. Parameter telemetry
160
+ omits positional arguments, raw command arguments, paths, IDs, and arbitrary free-form values. Set
141
161
  `AGENTBRICKS_DISABLE_TELEMETRY=1` to disable these uploads.
142
162
 
143
163
  Telemetry runs only when a leaf command is invoked, whether its callback succeeds or fails. Help
@@ -150,6 +170,25 @@ thread, returning sooner if the worker finishes. This is not a delivery guarante
150
170
  uploads can be lost when the process exits after the allowance expires. Missing credentials or a
151
171
  failed upload never change the command result.
152
172
 
173
+ Deploy and `deployments wait` telemetry includes a random invocation UUID, whether a gated
174
+ agent-details link was printed, the observed deployment state, and whether the process is an
175
+ internal background child. Details links carry the invocation UUID so browser arrivals can be
176
+ joined to the command that printed them. The UUID contains no agent name or authentication data.
177
+ Rollout outcomes are recorded independently of link availability, including failed rollouts;
178
+ preparation failures leave the deployment outcome unspecified.
179
+ `AGENTBRICKS_DISABLE_TELEMETRY=1` also removes these attribution parameters while keeping the link.
180
+ `deployments get` and `deployments list` also attach a per-command UUID to details links.
181
+ All rows from one list share one UUID and one command event; their deployment outcome stays
182
+ unspecified because these commands only read state. List links count as emitted only when JSON
183
+ output or terminal hyperlink rendering preserves their targets. Plain redirected output and
184
+ terminal modes that strip hyperlink targets keep `details_link_emitted=false`; use `--output json`
185
+ to capture the URLs. Group arrivals by the joined event's
186
+ `command_path` to distinguish deploy, wait, get, and list. Count distinct invocation IDs, rather
187
+ than list rows or repeated page loads, for command conversion percentages.
188
+ A wait command has its own invocation ID; it is not another deploy. Exclude background children
189
+ from user-invoked command counts. Link arrivals measure browser opens, including copied links,
190
+ not terminal mouse clicks or proof that the deployer and viewer are the same person.
191
+
153
192
  ## Quickstart
154
193
 
155
194
  The shortest path from a blank directory to a running and deployed agent:
@@ -185,6 +224,18 @@ The app's compute starts first and boots while the other resources are provision
185
224
  deploy stops partway, the app it created is kept: re-run `agentbricks deploy` to resume, or remove
186
225
  it with `agentbricks deployments delete agent-bricks-my-agent`.
187
226
 
227
+ When the workspace has both the Agent Memory Sessions beta and the new Agents UI enabled,
228
+ deploy output also includes an **Agent details** link to the workspace management page. This
229
+ applies to first deploys, redeploys, `--no-wait`, and `deployments wait`.
230
+ `deployments get` shows the same link alongside the app URL; `deployments list` links agent
231
+ names to their details pages and keeps a separate **Open app** link. JSON output adds
232
+ `details_url` when available and retains `url` for the deployed agent. The optional flag lookup
233
+ waits at most two seconds per command, including lists; unavailable or disabled flags omit the
234
+ link without affecting the command.
235
+ When following a local deploy record, `deployments wait` also requires its recorded workspace
236
+ host and ID to match the selected profile. Older records or records without a workspace ID omit
237
+ the management link; the recorded deployment status and invocation URL are still available.
238
+
188
239
  ### Deploying from a coding agent
189
240
 
190
241
  A first deploy takes ~3 min (mostly starting app compute); redeploys are faster. Either run
@@ -672,8 +723,8 @@ agentbricks memory stores permissions revoke memory-stores/<id> \
672
723
 
673
724
  Session stores accept their configured name or `session-stores/<name>`. Memory stores accept an ID
674
725
  or `memory-stores/<id>`; permission Get also accepts the display-name lookup supported by store Get
675
- and rejects ambiguous matches. The global `--profile` and `--output json` options go before
676
- `sessions` or `memory`. List follows the existing CLI pagination convention: one JSON/piped page
726
+ and rejects ambiguous matches. The global `--profile` option goes before `sessions` or `memory`;
727
+ `--output json` goes before or after the command. List follows the existing CLI pagination convention: one JSON/piped page
677
728
  with `next_page_token`, interactive text pages, or every page with `--all`.
678
729
 
679
730
  Get always returns the authenticated caller's effective permission, including
@@ -796,7 +847,10 @@ Use the permission commands below as the store creator or a workspace admin to r
796
847
  ## Model and prompt upgrades (experimental)
797
848
 
798
849
  `agentbricks experimental models` and `agentbricks experimental prompts` find the best model and
799
- prompt for each LLM call in your agent and swap them in. They live under `experimental` because
850
+ prompt for each LLM call in your agent and swap them in. Use them when a new model launches, or when
851
+ you want the same quality for less: you supply an eval set and how much quality, latency, and cost
852
+ matter, and the search tries each candidate model per LLM call, rewrites the prompts for it, and
853
+ recommends the combination that scores best on your whole agent. They live under `experimental` because
800
854
  their flags and output may still change between releases without a deprecation period. The
801
855
  agent.toml tables they write live under `[experimental]` too.
802
856
 
@@ -817,17 +871,56 @@ agentbricks experimental models status # the job's state, then its recomm
817
871
  agentbricks experimental models apply # switch to it
818
872
  ```
819
873
 
874
+ The `agent.eval:...` references name things in your project. A minimal `agent/eval.py`:
875
+
876
+ ```python
877
+ import mlflow
878
+
879
+ from agent.agent import answer # your agent's own entry point, called synchronously
880
+
881
+ mlflow.langchain.autolog() # traces token usage, which cost scoring needs; use your framework's autolog
882
+
883
+
884
+ def predict(inputs):
885
+ return answer(inputs["question"]) # wrap an async agent in asyncio.run(...)
886
+
887
+
888
+ TRAIN = [
889
+ {
890
+ "inputs": {"question": "Plan 3 days in Tokyo in late March"},
891
+ "expectations": {"expected_response": "..."},
892
+ },
893
+ # ... a few dozen records that look like real traffic
894
+ ]
895
+ VAL = [...] # same shape, held out; the winner and the summary's medians come from these
896
+
897
+
898
+ def mentions_expected(inputs, expectations, answer):
899
+ return float(expectations["expected_response"].lower() in str(answer).lower())
900
+
901
+
902
+ SCORERS = [mentions_expected] # or MLflow scorers, e.g. mlflow.genai.scorers.Correctness()
903
+ ```
904
+
905
+ Each of the agent's LLM calls must call its role's model service, `resolve_model_service("<role>")`
906
+ from `databricks_agentkit.runtime.tool_manifest`, so the search can route it to each candidate.
907
+
820
908
  By default `models upgrade` uploads the project to your workspace and runs the search as a
821
909
  serverless Databricks job, because a search can take hours. `--run-on local` runs it in your
822
910
  terminal instead, which needs the `upgrade` extra and the agent's dependencies installed. Each call
823
911
  takes longer from a laptop than on a serverless job, and a record slower than `--latency-gate` (60
824
912
  seconds by default) scores 0, so raise it if a local run warns that records went over the gate. The job imports the `predict_fn`,
825
913
  eval data, and scorers you name from the project (the same inputs `optimize_prompts_and_models`
826
- takes, below) and runs each eval record once per candidate, against a temporary `<service>_exp_<id>`
827
- clone. Each LLM call must use its role's service, `resolve_model_service("<role>")`, which reads the
914
+ takes, below) and runs each eval record once per candidate, against temporary clones of the model
915
+ services: one per service and model it may use, `<service>_exp_<id>_<model>`, created up front and
916
+ deleted when the search ends. Each clone always serves its one model, so switching candidates
917
+ redirects calls to another clone instead of waiting for a service's model to change. Each LLM call must use its role's service, `resolve_model_service("<role>")`, which reads the
828
918
  `AGENT_MODEL_SERVICE_<ROLE>` env var that deploy (and the job, before importing your code) sets.
829
- GEPA also rewrites every prompt bound with `prompts bind` for each candidate model. Pass
830
- `--prompt` to pick specific prompts instead, or `--models-only` to leave prompts alone. `models apply` switches
919
+ GEPA also rewrites every prompt bound with `prompts bind` for each candidate model, so you don't
920
+ pass prompt URIs unless you want a different set. Pass `--prompt` to pick specific prompts instead,
921
+ or `--models-only` to leave prompts alone. A `--prompt` URI must name an alias
922
+ (`prompts:/main.my_agent.writer@production`), not a version (`.../writer/3`): apply moves the
923
+ alias to the rewritten version, so an agent that loads a fixed version would never see it. `models apply` switches
831
924
  to the best quality / latency / cost trade-off and registers any rewritten prompts (prior versions
832
925
  keep `@production_previous`). `models rollback` undoes the whole last apply: it switches every
833
926
  model back and moves each prompt's alias back to its prior version. Apply refuses recommendations
@@ -836,6 +929,11 @@ recorded when the recommendation was applied. If rollback fails partway through,
836
929
  to finish the remaining changes. Each search owns its temporary clones, so overlapping jobs can
837
930
  evaluate different models independently.
838
931
 
932
+ When the search finishes, the summary shows each role's current and recommended model, the change in
933
+ the combined score, and how the median quality, latency, and cost per request moved from the
934
+ current setup to the recommendation on the validation records, for example
935
+ `Median cost per request $0.0040 → $0.0010 (-75%)`. `models status` shows the same summary later.
936
+
839
937
  The search is `databricks_agentkit.model_upgrades`, which you can also call directly (for example
840
938
  from a notebook) to tune several model services and MLflow Prompt Registry prompts together:
841
939
 
@@ -1085,7 +1183,7 @@ requests. By default, each App user's `X-Routing-Key` matches its generated user
1085
1183
 
1086
1184
  The final summary includes counts, throughput including drain, successful-only and all-completed
1087
1185
  latency percentiles, answer-text TTFT, token usage, coverage, incomplete attempts, and generator CPU
1088
- warnings. With `--output json` before `endpoint`, stdout contains one Agent Bricks summary object;
1186
+ warnings. With `--output json`, stdout contains one Agent Bricks summary object;
1089
1187
  progress stays on stderr. This is the same object saved as `summary.json`.
1090
1188
  Locust's rounded histogram buckets are used for percentiles; all-completed timings include failures.
1091
1189
  These bounded tests measure the selected workload and do not establish sustainable capacity or answer quality.
@@ -1471,9 +1569,8 @@ agentbricks tools remove genie_agent
1471
1569
  to `user`; choose `--auth app` deliberately for App service-principal execution. Existing manifests
1472
1570
  without `auth` preserve App/default identity. Both add commands and `remove` accept `--source PATH`
1473
1571
  to select a project instead of the current directory. Discovery needs no project; read that
1474
- project's `agent.toml` to inspect configured bindings. For scripted output, put the global
1475
- `-o json` option before `tools`, as in
1476
- `agentbricks -o json tools add genie-one --source ./my-agent`. Adding a binding is offline: it updates
1572
+ project's `agent.toml` to inspect configured bindings. For scripted output, add `-o json`, as in
1573
+ `agentbricks tools add genie-one --source ./my-agent -o json`. Adding a binding is offline: it updates
1477
1574
  `agent.toml` without contacting Genie or checking permissions. The corresponding sources are:
1478
1575
 
1479
1576
  ```toml
@@ -1565,3 +1662,20 @@ sync/streaming/background transport selector is manual.
1565
1662
  Developing Agent Bricks CLI (`agentbricks`), AgentKit, the runtime, and templates - plus the local dev loop and how to
1566
1663
  test unreleased changes on `agentbricks dev` and `agentbricks deploy`, is covered in
1567
1664
  [CONTRIBUTING.md](CONTRIBUTING.md).
1665
+
1666
+ The deployment workflow tracks phases and readiness internally; these observations are not
1667
+ logged as separate telemetry fields. Network errors are retried within the existing readiness budget; if the final probe remains unreachable,
1668
+ the deployed app has unverified readiness rather than a reported application failure. These
1669
+ observations do not validate model or tool behavior.
1670
+
1671
+ Command outcomes can be interpreted from the existing command path, exit code, error category,
1672
+ and options. Pending waits (`deployments wait` exit 75) and background submissions
1673
+ (`deploy --no-wait`) should not be counted as deployment failures. Deployment operation UUIDs
1674
+ stay in local state and the background worker; they are not uploaded. The separate uploaded
1675
+ `invocation_id` identifies one command and joins details-link arrivals to it.
1676
+ See [outcome interpretation](CONTRIBUTING.md#interpreting-command-and-deployment-outcomes) for
1677
+ failure-rate denominators and coverage limits.
1678
+
1679
+ Local runs track parent interruption and child exit status internally. A child-only SIGINT or
1680
+ explicit exit 130 remains a runtime failure; it is not treated as a confirmed user cancellation.
1681
+ Caught parent interruptions use the existing interruption category and omit failure text.
@@ -1,6 +1,6 @@
1
1
  [project]
2
2
  name = "databricks-agentbricks"
3
- version = "0.9.0"
3
+ version = "0.11.0"
4
4
  description = "Agent Bricks CLI and AgentKit SDK for custom agents on Databricks"
5
5
  authors = [
6
6
  { name="Databricks", email="agent-feedback@databricks.com" },
@@ -123,9 +123,9 @@ def _emit_trailing_spacer() -> None:
123
123
  try:
124
124
  # Suppress under `-o json` even on a (pseudo-)TTY — a machine may be parsing stdout under
125
125
  # `docker run -t` / `script` / a CI PTY, where isatty() alone would wrongly allow the
126
- # spacer. `errors._OUTPUT_MODE` is a process global (set by the root group, read the same
127
- # way by render.py), so it's still available here after the Click context is torn down.
128
- if errors._OUTPUT_MODE == "json" or not sys.stdout.isatty():
126
+ # spacer. `errors.error_output_mode()` is a process global (set by the root group, read the
127
+ # same way by render.py), so it's still available here after the Click context is torn down.
128
+ if errors.error_output_mode() == "json" or not sys.stdout.isatty():
129
129
  return
130
130
  click.echo()
131
131
  except Exception: