generic-ml-wrapper 0.3.0__tar.gz → 0.5.0__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 (308) hide show
  1. generic_ml_wrapper-0.5.0/.github/workflows/release.yml +47 -0
  2. generic_ml_wrapper-0.5.0/CHANGELOG.md +231 -0
  3. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/PKG-INFO +3 -1
  4. generic_ml_wrapper-0.5.0/ROADMAP.md +280 -0
  5. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/SECURITY.md +1 -1
  6. generic_ml_wrapper-0.5.0/VERSION +1 -0
  7. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/docs/CLI.md +110 -0
  8. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/docs/CONFIGURATION.md +7 -3
  9. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/docs/DESIGN.md +4 -2
  10. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/docs/USER_GUIDE.md +19 -5
  11. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/docs/WORKFLOWS.md +2 -1
  12. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/hatch_build.py +7 -20
  13. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/pyproject.toml +12 -1
  14. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/__init__.py +1 -1
  15. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/adapter/inbound/cli/app.py +480 -110
  16. generic_ml_wrapper-0.5.0/src/generic_ml_wrapper/adapter/inbound/cli/help_topics.py +57 -0
  17. generic_ml_wrapper-0.5.0/src/generic_ml_wrapper/adapter/inbound/cli/hints.py +66 -0
  18. generic_ml_wrapper-0.5.0/src/generic_ml_wrapper/adapter/inbound/cli/index.py +71 -0
  19. generic_ml_wrapper-0.5.0/src/generic_ml_wrapper/adapter/outbound/bootstrap/filesystem_layout_migrator.py +61 -0
  20. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/adapter/outbound/bootstrap/filesystem_layout_seeder.py +199 -13
  21. generic_ml_wrapper-0.5.0/src/generic_ml_wrapper/adapter/outbound/bootstrap/http_client_versions.py +169 -0
  22. generic_ml_wrapper-0.5.0/src/generic_ml_wrapper/adapter/outbound/bootstrap/subprocess_command_runner.py +33 -0
  23. generic_ml_wrapper-0.5.0/src/generic_ml_wrapper/adapter/outbound/bootstrap/system_clipboard.py +57 -0
  24. generic_ml_wrapper-0.5.0/src/generic_ml_wrapper/adapter/outbound/bootstrap/tty_client_setup.py +246 -0
  25. generic_ml_wrapper-0.5.0/src/generic_ml_wrapper/adapter/outbound/bootstrap/tty_language_chooser.py +54 -0
  26. generic_ml_wrapper-0.5.0/src/generic_ml_wrapper/adapter/outbound/bootstrap/tty_persona_chooser.py +51 -0
  27. generic_ml_wrapper-0.5.0/src/generic_ml_wrapper/adapter/outbound/bootstrap/tty_prompt.py +120 -0
  28. generic_ml_wrapper-0.5.0/src/generic_ml_wrapper/adapter/outbound/bootstrap/tty_text_prompt.py +56 -0
  29. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/adapter/outbound/caller/claude_cli_caller.py +2 -1
  30. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/adapter/outbound/caller/codex_cli_caller.py +2 -1
  31. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/adapter/outbound/caller/status_line_config.py +3 -2
  32. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/adapter/outbound/caller/vibe_cli_caller.py +4 -3
  33. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/adapter/outbound/compress/cache_backed_compressor.py +5 -3
  34. generic_ml_wrapper-0.5.0/src/generic_ml_wrapper/adapter/outbound/config/__init__.py +3 -0
  35. generic_ml_wrapper-0.5.0/src/generic_ml_wrapper/adapter/outbound/config/tomlkit_config_writer.py +79 -0
  36. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/adapter/outbound/gateway/relay.py +2 -1
  37. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/adapter/outbound/interceptor/size_logger.py +2 -1
  38. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/adapter/outbound/workflow/filesystem_workflow_source.py +59 -8
  39. generic_ml_wrapper-0.5.0/src/generic_ml_wrapper/application/domain/model/client_catalog.py +224 -0
  40. generic_ml_wrapper-0.5.0/src/generic_ml_wrapper/application/domain/model/migration.py +33 -0
  41. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/domain/service/greeting.py +21 -0
  42. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/domain/service/hook_runner.py +2 -1
  43. generic_ml_wrapper-0.5.0/src/generic_ml_wrapper/application/port/inbound/config_commands.py +84 -0
  44. generic_ml_wrapper-0.5.0/src/generic_ml_wrapper/application/port/inbound/edit_workflow.py +44 -0
  45. generic_ml_wrapper-0.5.0/src/generic_ml_wrapper/application/port/inbound/init.py +50 -0
  46. generic_ml_wrapper-0.5.0/src/generic_ml_wrapper/application/port/inbound/migrate_layout.py +23 -0
  47. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/port/inbound/start_job.py +17 -2
  48. generic_ml_wrapper-0.5.0/src/generic_ml_wrapper/application/port/outbound/client_setup.py +35 -0
  49. generic_ml_wrapper-0.5.0/src/generic_ml_wrapper/application/port/outbound/client_version.py +46 -0
  50. generic_ml_wrapper-0.5.0/src/generic_ml_wrapper/application/port/outbound/clipboard.py +23 -0
  51. generic_ml_wrapper-0.5.0/src/generic_ml_wrapper/application/port/outbound/command_runner.py +22 -0
  52. generic_ml_wrapper-0.5.0/src/generic_ml_wrapper/application/port/outbound/config_writer.py +41 -0
  53. generic_ml_wrapper-0.5.0/src/generic_ml_wrapper/application/port/outbound/language_chooser.py +28 -0
  54. generic_ml_wrapper-0.5.0/src/generic_ml_wrapper/application/port/outbound/layout_migrator.py +31 -0
  55. generic_ml_wrapper-0.5.0/src/generic_ml_wrapper/application/port/outbound/layout_seeder.py +85 -0
  56. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/port/outbound/persona_chooser.py +5 -1
  57. generic_ml_wrapper-0.5.0/src/generic_ml_wrapper/application/port/outbound/text_prompt.py +32 -0
  58. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/port/outbound/workflow_source.py +11 -0
  59. generic_ml_wrapper-0.5.0/src/generic_ml_wrapper/application/usecase/edit_workflow.py +112 -0
  60. generic_ml_wrapper-0.5.0/src/generic_ml_wrapper/application/usecase/init.py +126 -0
  61. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/usecase/launch.py +2 -1
  62. generic_ml_wrapper-0.5.0/src/generic_ml_wrapper/application/usecase/migrate_layout.py +42 -0
  63. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/usecase/start_job.py +45 -3
  64. generic_ml_wrapper-0.5.0/src/generic_ml_wrapper/application/usecase/update_config.py +86 -0
  65. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/wiring/composition.py +127 -9
  66. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/common/config.py +140 -7
  67. generic_ml_wrapper-0.5.0/src/generic_ml_wrapper/common/i18n.py +146 -0
  68. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/common/paths.py +5 -0
  69. generic_ml_wrapper-0.5.0/src/generic_ml_wrapper/common/settings_registry.py +371 -0
  70. generic_ml_wrapper-0.5.0/src/generic_ml_wrapper/resources/i18n/en.json +145 -0
  71. generic_ml_wrapper-0.5.0/src/generic_ml_wrapper/resources/i18n/fr.json +145 -0
  72. generic_ml_wrapper-0.5.0/tests/test_bootstrap.py +205 -0
  73. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_check_client_ready.py +4 -1
  74. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_cli.py +402 -88
  75. generic_ml_wrapper-0.5.0/tests/test_client_catalog.py +43 -0
  76. generic_ml_wrapper-0.5.0/tests/test_client_versions.py +99 -0
  77. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_config.py +25 -0
  78. generic_ml_wrapper-0.5.0/tests/test_edit_workflow_usecase.py +123 -0
  79. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_exit_handling.py +0 -6
  80. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_filesystem_workflow_source.py +85 -12
  81. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_greeting.py +7 -0
  82. generic_ml_wrapper-0.5.0/tests/test_i18n.py +107 -0
  83. generic_ml_wrapper-0.5.0/tests/test_init_usecase.py +196 -0
  84. generic_ml_wrapper-0.5.0/tests/test_layout_migrator.py +84 -0
  85. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_list_workflows_usecase.py +3 -0
  86. generic_ml_wrapper-0.5.0/tests/test_migrate_layout_usecase.py +32 -0
  87. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_new_workflow_usecase.py +3 -0
  88. generic_ml_wrapper-0.5.0/tests/test_settings_registry.py +115 -0
  89. generic_ml_wrapper-0.5.0/tests/test_smoke.py +28 -0
  90. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_start_job_usecase.py +66 -3
  91. generic_ml_wrapper-0.5.0/tests/test_tty_client_setup.py +204 -0
  92. generic_ml_wrapper-0.5.0/tests/test_tty_language_chooser.py +60 -0
  93. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_tty_persona_chooser.py +11 -9
  94. generic_ml_wrapper-0.5.0/tests/test_tty_prompt.py +86 -0
  95. generic_ml_wrapper-0.5.0/tests/test_tty_text_prompt.py +62 -0
  96. generic_ml_wrapper-0.5.0/tests/test_update_config_usecase.py +86 -0
  97. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/uv.lock +175 -1
  98. generic_ml_wrapper-0.3.0/.github/workflows/release.yml +0 -28
  99. generic_ml_wrapper-0.3.0/CHANGELOG.md +0 -110
  100. generic_ml_wrapper-0.3.0/ROADMAP.md +0 -173
  101. generic_ml_wrapper-0.3.0/VERSION +0 -1
  102. generic_ml_wrapper-0.3.0/src/generic_ml_wrapper/adapter/outbound/bootstrap/tty_client_chooser.py +0 -49
  103. generic_ml_wrapper-0.3.0/src/generic_ml_wrapper/adapter/outbound/bootstrap/tty_persona_chooser.py +0 -52
  104. generic_ml_wrapper-0.3.0/src/generic_ml_wrapper/application/domain/model/client_catalog.py +0 -77
  105. generic_ml_wrapper-0.3.0/src/generic_ml_wrapper/application/port/inbound/first_run_init.py +0 -37
  106. generic_ml_wrapper-0.3.0/src/generic_ml_wrapper/application/port/outbound/client_chooser.py +0 -23
  107. generic_ml_wrapper-0.3.0/src/generic_ml_wrapper/application/port/outbound/layout_seeder.py +0 -25
  108. generic_ml_wrapper-0.3.0/src/generic_ml_wrapper/application/usecase/first_run_init.py +0 -71
  109. generic_ml_wrapper-0.3.0/tests/test_bootstrap.py +0 -101
  110. generic_ml_wrapper-0.3.0/tests/test_first_run_init.py +0 -123
  111. generic_ml_wrapper-0.3.0/tests/test_smoke.py +0 -20
  112. generic_ml_wrapper-0.3.0/tests/test_tty_client_chooser.py +0 -55
  113. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/.coveragerc +0 -0
  114. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/.gitattributes +0 -0
  115. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/.github/workflows/ci.yml +0 -0
  116. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/.github/workflows/pr-hygiene.yml +0 -0
  117. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/.github/workflows/sonar.yml +0 -0
  118. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/.gitignore +0 -0
  119. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/.importlinter +0 -0
  120. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/.pre-commit-config.yaml +0 -0
  121. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/AGENTS.md +0 -0
  122. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/CODE_OF_CONDUCT.md +0 -0
  123. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/CONTRIBUTING.md +0 -0
  124. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/GOVERNANCE.md +0 -0
  125. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/LICENSE +0 -0
  126. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/NOTICE +0 -0
  127. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/README.md +0 -0
  128. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/RELEASE.md +0 -0
  129. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/docs/CLIENTS.md +0 -0
  130. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/docs/TROUBLESHOOTING.md +0 -0
  131. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/docs/images/gmlw-favicon.svg +0 -0
  132. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/docs/images/gmlw-lockup-dark.svg +0 -0
  133. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/docs/images/gmlw-lockup.svg +0 -0
  134. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/docs/images/gmlw-statusline.gif +0 -0
  135. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/docs/images/gmlw-usage.gif +0 -0
  136. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/docs/tapes/README.md +0 -0
  137. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/docs/tapes/render.sh +0 -0
  138. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/docs/tapes/seed.py +0 -0
  139. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/docs/tapes/statusline.tape +0 -0
  140. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/docs/tapes/usage.tape +0 -0
  141. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/noxfile.py +0 -0
  142. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/pyrightconfig.json +0 -0
  143. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/sonar-project.properties +0 -0
  144. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/adapter/__init__.py +0 -0
  145. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/adapter/inbound/__init__.py +0 -0
  146. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/adapter/inbound/cli/__init__.py +0 -0
  147. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/adapter/inbound/cli/banner.py +0 -0
  148. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/adapter/outbound/__init__.py +0 -0
  149. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/adapter/outbound/bootstrap/__init__.py +0 -0
  150. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/adapter/outbound/bootstrap/path_client_detector.py +0 -0
  151. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/adapter/outbound/caller/__init__.py +0 -0
  152. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/adapter/outbound/caller/context_file.py +0 -0
  153. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/adapter/outbound/caller/context_opening.py +0 -0
  154. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/adapter/outbound/caller/cursor_cli_caller.py +0 -0
  155. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/adapter/outbound/caller/default_provider.py +0 -0
  156. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/adapter/outbound/caller/loader.py +0 -0
  157. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/adapter/outbound/caller/vibe_config.py +0 -0
  158. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/adapter/outbound/compress/__init__.py +0 -0
  159. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/adapter/outbound/credentials/__init__.py +0 -0
  160. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/adapter/outbound/credentials/filesystem_credentials_store.py +0 -0
  161. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/adapter/outbound/gateway/__init__.py +0 -0
  162. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/adapter/outbound/gateway/anthropic_sse.py +0 -0
  163. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/adapter/outbound/gateway/openai_chat.py +0 -0
  164. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/adapter/outbound/gateway/openai_responses.py +0 -0
  165. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/adapter/outbound/hook/__init__.py +0 -0
  166. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/adapter/outbound/hook/session_logger.py +0 -0
  167. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/adapter/outbound/interceptor/__init__.py +0 -0
  168. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/adapter/outbound/persona/__init__.py +0 -0
  169. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/adapter/outbound/persona/filesystem_persona_source.py +0 -0
  170. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/adapter/outbound/plugin/__init__.py +0 -0
  171. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/adapter/outbound/plugin/filesystem_plugin_source.py +0 -0
  172. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/adapter/outbound/status/__init__.py +0 -0
  173. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/adapter/outbound/status/claude_status_parser.py +0 -0
  174. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/adapter/outbound/status/cursor_status_parser.py +0 -0
  175. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/adapter/outbound/store/__init__.py +0 -0
  176. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/adapter/outbound/store/filesystem_transcript_store.py +0 -0
  177. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/adapter/outbound/store/ledger.py +0 -0
  178. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/adapter/outbound/store/sqlite_per_turn_store.py +0 -0
  179. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/adapter/outbound/store/sqlite_session_store.py +0 -0
  180. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/adapter/outbound/store/sqlite_usage_store.py +0 -0
  181. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/adapter/outbound/workflow/__init__.py +0 -0
  182. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/adapter/outbound/workspace/__init__.py +0 -0
  183. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/adapter/outbound/workspace/local_workspace_inspector.py +0 -0
  184. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/__init__.py +0 -0
  185. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/domain/__init__.py +0 -0
  186. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/domain/model/__init__.py +0 -0
  187. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/domain/model/client_status.py +0 -0
  188. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/domain/model/context_source.py +0 -0
  189. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/domain/model/identifiers.py +0 -0
  190. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/domain/model/learned.py +0 -0
  191. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/domain/model/persona.py +0 -0
  192. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/domain/model/plugin.py +0 -0
  193. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/domain/model/rules.py +0 -0
  194. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/domain/model/run.py +0 -0
  195. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/domain/model/session.py +0 -0
  196. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/domain/model/turn_usage.py +0 -0
  197. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/domain/model/workspace.py +0 -0
  198. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/domain/service/__init__.py +0 -0
  199. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/domain/service/hook.py +0 -0
  200. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/domain/service/interceptor.py +0 -0
  201. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/domain/service/interceptor_chain.py +0 -0
  202. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/domain/service/persona_parser.py +0 -0
  203. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/domain/service/rule_cleaner.py +0 -0
  204. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/domain/service/session_naming.py +0 -0
  205. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/domain/service/statusline_renderer.py +0 -0
  206. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/port/__init__.py +0 -0
  207. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/port/inbound/__init__.py +0 -0
  208. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/port/inbound/bootstrap.py +0 -0
  209. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/port/inbound/check_client_ready.py +0 -0
  210. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/port/inbound/export_usage.py +0 -0
  211. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/port/inbound/list_jobs.py +0 -0
  212. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/port/inbound/list_personas.py +0 -0
  213. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/port/inbound/list_plugins.py +0 -0
  214. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/port/inbound/list_sessions.py +0 -0
  215. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/port/inbound/list_workflows.py +0 -0
  216. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/port/inbound/new_workflow.py +0 -0
  217. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/port/inbound/render_greeting.py +0 -0
  218. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/port/inbound/render_statusline.py +0 -0
  219. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/port/inbound/set_credential.py +0 -0
  220. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/port/outbound/__init__.py +0 -0
  221. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/port/outbound/cli_caller.py +0 -0
  222. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/port/outbound/client_detector.py +0 -0
  223. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/port/outbound/client_status.py +0 -0
  224. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/port/outbound/context_compressor.py +0 -0
  225. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/port/outbound/credentials_store.py +0 -0
  226. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/port/outbound/hook.py +0 -0
  227. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/port/outbound/interceptor.py +0 -0
  228. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/port/outbound/per_turn_metering.py +0 -0
  229. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/port/outbound/persona_source.py +0 -0
  230. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/port/outbound/plugin_source.py +0 -0
  231. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/port/outbound/session_store.py +0 -0
  232. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/port/outbound/transcript.py +0 -0
  233. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/port/outbound/usage_store.py +0 -0
  234. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/port/outbound/workspace.py +0 -0
  235. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/usecase/__init__.py +0 -0
  236. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/usecase/bootstrap.py +0 -0
  237. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/usecase/check_client_ready.py +0 -0
  238. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/usecase/export_usage.py +0 -0
  239. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/usecase/list_jobs.py +0 -0
  240. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/usecase/list_personas.py +0 -0
  241. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/usecase/list_plugins.py +0 -0
  242. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/usecase/list_sessions.py +0 -0
  243. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/usecase/list_workflows.py +0 -0
  244. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/usecase/new_workflow.py +0 -0
  245. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/usecase/render_greeting.py +0 -0
  246. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/usecase/render_statusline.py +0 -0
  247. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/usecase/set_credential.py +0 -0
  248. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/application/wiring/__init__.py +0 -0
  249. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/common/__init__.py +0 -0
  250. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/common/log.py +0 -0
  251. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/common/spec_loader.py +0 -0
  252. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/py.typed +0 -0
  253. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/resources/personas/_floor.md +0 -0
  254. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/resources/personas/butler.md +0 -0
  255. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/resources/personas/companion.md +0 -0
  256. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/resources/personas/mentor.md +0 -0
  257. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/resources/personas/plain.md +0 -0
  258. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/resources/personas/terse.md +0 -0
  259. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/resources/workflows/_common/base.md +0 -0
  260. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/src/generic_ml_wrapper/resources/workflows/create-workflow/workflow.md +0 -0
  261. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/_conformance.py +0 -0
  262. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/conftest.py +0 -0
  263. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_anthropic_sse.py +0 -0
  264. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_authoring_separation.py +0 -0
  265. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_banner.py +0 -0
  266. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_cache_backed_compressor.py +0 -0
  267. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_caller_loader.py +0 -0
  268. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_callers.py +0 -0
  269. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_context_file.py +0 -0
  270. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_context_source.py +0 -0
  271. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_credentials_store.py +0 -0
  272. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_cursor_status_parser.py +0 -0
  273. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_docs_consistency.py +0 -0
  274. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_export_usage_usecase.py +0 -0
  275. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_filesystem_persona_source.py +0 -0
  276. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_filesystem_plugin_source.py +0 -0
  277. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_filesystem_transcript_store.py +0 -0
  278. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_hook_runner.py +0 -0
  279. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_identifiers.py +0 -0
  280. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_interceptor_chain.py +0 -0
  281. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_learned.py +0 -0
  282. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_list_jobs_usecase.py +0 -0
  283. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_list_personas_usecase.py +0 -0
  284. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_list_plugins_usecase.py +0 -0
  285. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_list_sessions_usecase.py +0 -0
  286. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_local_workspace_inspector.py +0 -0
  287. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_log.py +0 -0
  288. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_model_invariants.py +0 -0
  289. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_openai_chat.py +0 -0
  290. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_openai_responses.py +0 -0
  291. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_path_client_detector.py +0 -0
  292. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_persona_parser.py +0 -0
  293. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_relay.py +0 -0
  294. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_render_greeting_usecase.py +0 -0
  295. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_rule_cleaner.py +0 -0
  296. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_session_logger.py +0 -0
  297. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_session_naming.py +0 -0
  298. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_size_logger.py +0 -0
  299. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_sqlite_per_turn_store.py +0 -0
  300. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_sqlite_session_store.py +0 -0
  301. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_sqlite_usage_store.py +0 -0
  302. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_status_line_config.py +0 -0
  303. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_statusline.py +0 -0
  304. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_store_conformance.py +0 -0
  305. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_version.py +0 -0
  306. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tests/test_vibe_config.py +0 -0
  307. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tools/hooks/check-commit-msg.sh +0 -0
  308. {generic_ml_wrapper-0.3.0 → generic_ml_wrapper-0.5.0}/tools/hooks/guard-branch.sh +0 -0
@@ -0,0 +1,47 @@
1
+ # Publish to PyPI via Trusted Publishing (OIDC) on a version tag (e.g. v0.5.0). No token
2
+ # or secret: PyPI trusts this exact repo + workflow + environment, and GitHub mints a
3
+ # short-lived OIDC credential for the upload. Release flow (matches generic-ml-cache):
4
+ # bump VERSION + CHANGELOG on a release branch, merge to main, then tag the head. Every
5
+ # build-time input is pinned (Actions by commit SHA, uv by [tool.uv] required-version).
6
+ name: release
7
+
8
+ on:
9
+ push:
10
+ tags: ["v*"]
11
+
12
+ permissions:
13
+ contents: read
14
+
15
+ jobs:
16
+ # Fail fast if the tag does not match VERSION, so a mistagged release (e.g. v0.5.0
17
+ # pointing at a commit still on 0.4.0) never builds and publishes the wrong version.
18
+ guard:
19
+ name: Tag matches version
20
+ runs-on: ubuntu-latest
21
+ steps:
22
+ - uses: actions/checkout@df4cb1c069e1874edd31b4311f1884172cec0e10 # v6
23
+ - name: Tag must equal VERSION
24
+ run: |
25
+ tag="${GITHUB_REF_NAME#v}"
26
+ version="$(tr -d '[:space:]' < VERSION)"
27
+ if [ "$version" != "$tag" ]; then
28
+ echo "::error::tag ${GITHUB_REF_NAME} does not match VERSION ($version)"
29
+ exit 1
30
+ fi
31
+ echo "tag v$tag matches VERSION $version"
32
+
33
+ pypi-publish:
34
+ name: Build and publish to PyPI
35
+ needs: guard
36
+ runs-on: ubuntu-latest
37
+ environment: pypi
38
+ permissions:
39
+ id-token: write # OIDC token for Trusted Publishing
40
+ contents: read
41
+ steps:
42
+ - uses: actions/checkout@df4cb1c069e1874edd31b4311f1884172cec0e10 # v6
43
+ - uses: astral-sh/setup-uv@11f9893b081a58869d3b5fccaea48c9e9e46f990 # v8.3.2
44
+ - name: Build sdist + wheel
45
+ run: uv build
46
+ - name: Publish to PyPI (Trusted Publishing)
47
+ run: uv publish --trusted-publishing always
@@ -0,0 +1,231 @@
1
+ # Changelog
2
+
3
+ All notable changes to this project are documented here. The format follows
4
+ [Keep a Changelog](https://keepachangelog.com/en/1.1.0/), and the project adheres
5
+ to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
6
+
7
+ ## [Unreleased]
8
+
9
+ ## [0.5.0] - 2026-07-19
10
+
11
+ Discoverability and progressive disclosure — shortening time-to-first-session and
12
+ revealing depth gradually — plus app-wide localisation.
13
+
14
+ ### Added
15
+ - **App-wide localisation.** Every user-facing message **and every diagnostic log line**
16
+ now renders through a process-global active localiser (`i18n.set_active` / `active` /
17
+ `t`), bound once at startup from the configured language, English-fallback then raw-key
18
+ safe. The EN/FR catalogues are kept in lockstep by a drift-guard test. Localising the
19
+ logs too is a deliberate choice, not the usual English-only-logs convention.
20
+ - **Config registry + `config` commands.** A `pydantic-settings` model is the typed source
21
+ of truth for every settable scalar key (type, default, allowed values, description).
22
+ `gmlw config list / get / set` render and validate against it; `set` merges through a
23
+ shared tomlkit writer (comments and formatting preserved, never rewritten) and surfaces
24
+ the old→new change. The home for changing `default_role` / `default_environment` after
25
+ init. Scope is the scalar keys; the structural matrices (`[[hooks]]`, `[[interceptors]]`,
26
+ `[startup.*.context]`, `[compress.prompts]`) stay hand-rolled, a deferred follow-up.
27
+ - **Bare `gmlw` capability index + `gmlw help <topic>`.** Bare `gmlw` is first-run-aware: a
28
+ fresh install runs `init`, thereafter it shows a grouped capability index
29
+ (launch / inspect / author) with a next-action footer. `gmlw help` explains the core
30
+ concepts (`job-vs-workflow`, `start-vs-run`, `personas`, `cost`). `--help` keeps the
31
+ argparse view.
32
+ - **Exit receipt.** On the return (client exit), a persistent summary: this session's and
33
+ the job's cost, the resume/report commands, and one usage-driven, suppressible tip (shown
34
+ once each; `[hints] show = false` disables). `StartJob` now returns a `StartJobResult` so
35
+ the receipt can name the session.
36
+ - **Ambient capability card.** An off-by-default context injection
37
+ (`[ambient] capability_card`): a localised "how do I … in gmlw" card appended to a new
38
+ session's context so the client can answer gmlw questions mid-session.
39
+ - **`gmlw workflow edit`.** Amend an existing workflow in an authoring session — opens its
40
+ folder, never creates or overwrites; an unknown workflow exits non-zero with guidance.
41
+
42
+ ### Changed
43
+ - **Greeting → context.** The launch-time host greeting (structurally invisible once the
44
+ client clears the screen) is no longer printed to stderr; it is injected into a new
45
+ session's context so the client renders it in-band. The parting `Bye, <name>.` on exit
46
+ stays.
47
+ - **`--version`.** Reports `gmlw <version> (build <id>)`; the git sha was dropped — it was
48
+ captured at build time and every distributed artifact is built without a `.git` checkout,
49
+ so it was always `unknown` where it would matter.
50
+
51
+ ### Notes
52
+ - Authoring-cost visibility is deferred: the authoring bucket is deliberately kept out of
53
+ `gmlw jobs`; surfacing its spend cleanly is its own design.
54
+
55
+ ## [0.4.0] - 2026-07-19
56
+
57
+ The first-run release — a mandatory `init` that establishes the model the rest of the app
58
+ runs on, migrates the on-disk layout to match, and settles a working client before the
59
+ first session.
60
+
61
+ ### Added
62
+ - **Guided client setup.** The init client step is no longer a silent chooser — it always
63
+ talks the choice through. It lists each installed client with its version, and when a
64
+ first-party release channel reports a newer one it flags an **old install** and offers
65
+ the one-line update (comparing on numeric components, so a build *ahead* of a lagging
66
+ stable channel is never nagged). It lets you switch, or **install a different client**:
67
+ it prints the OS-specific install command (macOS/Linux vs Windows), copies it to the
68
+ clipboard when a clipboard tool is present, offers to **run it for you** or let you run
69
+ it yourself, then polls `PATH` until the client appears — and installs a prerequisite
70
+ first (`uv` for Vibe) when it is missing. Every client's latest version comes from its
71
+ vendor's own channel with a changelog/registry fallback: Claude Code's native stable
72
+ manifest → GitHub `CHANGELOG.md`; Cursor's install-script version → the Homebrew cask
73
+ JSON; the npm registry → GitHub releases for Codex; PyPI → GitHub releases for Vibe.
74
+ All version reads are best-effort — an offline machine degrades to a plain list, never
75
+ a block. The launch-time "client not on your PATH" guidance now shows the same
76
+ OS-specific command. The client catalog (`client_catalog.py`) carries per-OS
77
+ install/update commands, the paid-plan framing, and the version sources.
78
+ - **Forced init + the gate.** A mandatory first-run setup, `gmlw init`, is now both a
79
+ command and a first-class gate: `[init] version` in `config.toml` records that it ran,
80
+ and any command on an un-initialised or pre-0.4.0 install is funnelled through init
81
+ before it runs (`statusline` and bare `--help` are exempt). The interview captures, in
82
+ order — each with a sensible default so a non-interactive run never blocks — **language
83
+ → name → role → environment → persona → client**: language sets the voice the rest of
84
+ the interview speaks (chosen language re-localises every later prompt), name is what the
85
+ companion calls you, role and environment seed the movie-set axes (`[profile]
86
+ default_role` / `default_environment`), persona and client reuse the existing choosers.
87
+ A fresh install gets a full seeded `config.toml`; a legacy install gets every answer
88
+ merged into its existing file (see below). Retires the thinner 0.2.0 `FirstRunInit`.
89
+ - **Every init answer is persisted — on a legacy install too.** Previously a pre-0.4.0
90
+ install had only the `[init]` marker appended, so the language, name, role, environment,
91
+ persona and client you had just chosen were discarded and had to be re-entered. They are
92
+ now **merged into the existing `config.toml`**: each value is written into its table
93
+ (created when missing) through a round-trip TOML edit, so **every other setting, every
94
+ comment, and the file's formatting survive untouched** — arrays like `[[interceptors]]`
95
+ included. The persona and client are written only when one was chosen, so declining never
96
+ clears an existing value. Any setting a fresh choice replaced is reported on stderr
97
+ (`client.default: cursor → claude`) rather than changed silently.
98
+ - **Environment migration.** Place-specific context is now a first-class **environment**:
99
+ it lives under `environments/<env>/` (one folder per environment, the movie set) instead
100
+ of the single `profile/company/`. On any command, gmlw non-destructively wraps an old
101
+ `profile/company/` into the **active** environment (`[profile] default_environment`,
102
+ `work` by default) — a move (nothing copied or lost), a name that already exists at the
103
+ target is left in place and reported (never overwritten), and the emptied old folder is
104
+ retired. What moved and what was skipped is printed to stderr. The move runs on both the
105
+ forced-init and the normal bootstrap paths, so an install initialised before the
106
+ migration existed is caught too. The `company` context source is unchanged as a config
107
+ key; only its on-disk home moved to `environments/<env>/`.
108
+ - **Role-scoped rules & learned.** The role chosen at init (`[profile] default_role`, a lens
109
+ over `me`) now shapes context: rules under `profile/roles/<role>/rules/*.md` and a
110
+ `profile/roles/<role>/learned.md` compose into the `rules` and `me.learned` sources **only
111
+ when that role is active** — layered after the global rules/learned (general → specific).
112
+ `gmlw init` seeds the chosen role's folder with an empty `rules/` drop-zone. Capture stays
113
+ global for now (new reflexes are still written to `rules/` and `profile/me/learned.md`);
114
+ role-aware capture is a later step.
115
+
116
+ ### Changed
117
+ - **New runtime dependency: `tomlkit`** (MIT, pure Python, no transitive deps) — the
118
+ round-trip TOML editor that merges init's answers into an existing `config.toml`
119
+ without disturbing the user's comments or settings. stdlib `tomllib` reads TOML but
120
+ cannot write it. `LayoutSeederPort.initialize` now returns an `InitPersist`
121
+ (`fresh` + `overwrites`) instead of a bare bool.
122
+ - **`ClientInfo` gained per-OS commands.** The single `install` field became
123
+ `install_unix` / `install_windows` (plus `update`, `subscription`, `version_probes`,
124
+ `prereq`); callers use `install_for(system)` / `update_for(system)`. The 0.2.0
125
+ `TtyClientChooser` and its `ClientChooserPort` were **removed**, superseded by the new
126
+ `ClientSetupPort` (`TtyClientSetup`) that owns the full guided conversation.
127
+
128
+ ## [0.3.0] - 2026-07-18
129
+
130
+ The lifecycle release — hooks that act at the seams around a run, and rule capture that
131
+ works everywhere.
132
+
133
+ ### Added
134
+ - **Lifecycle action hooks.** gmlw already hooks *content* (the interceptor chain); it now
135
+ hooks *actions* at two seams bracketing the client run. A `[[hooks]]` entry binds a
136
+ `HookPort` spec (a `module:Class` / `/path.py:Class`, or a plugin id) to a `phase` —
137
+ `pre-launch` (after the context is compiled and the caller resolved, before the client
138
+ starts) or `post-session` (after the client exits, with its exit code) — with an optional
139
+ `client` scope, under the same trusted-code boundary as `[[interceptors]]` and `[callers]`.
140
+ Hooks are best-effort: a failing hook never breaks a launch or its teardown. Both launch
141
+ paths (`start`, `workflow new`) route through one `run_with_hooks` sequence. Ships the
142
+ built-in `SessionLogger` as a reference hook.
143
+ - **Always-on rule lifecycle.** Rule capture is no longer workflow-only: the `rules`
144
+ context source (now active by default in a plain start, config-overridable) leads with
145
+ a capture directive so a demanded correction becomes a draft rule in **any** session.
146
+ The directive carries the full loop — offer to record a durable, reusable reflex;
147
+ **dedup** against the existing rules and update/supersede a match instead of stacking a
148
+ near-duplicate; and, when a rule is **mechanically enforceable**, offer to realise it as
149
+ a script or check rather than a standing reminder.
150
+
151
+ ### Fixed
152
+ - Docs: removed a stray tool artifact from `docs/CLIENTS.md`; corrected the workflow
153
+ compression default in `docs/CONFIGURATION.md` (every source defaults to
154
+ `compression = false`); documented `[companion] name`. Added a doc-consistency guard
155
+ against leaked tool-artifact tags.
156
+
157
+ ## [0.2.0] - 2026-07-16
158
+
159
+ The companion release — everything a session inherits, and the ergonomics around it.
160
+
161
+ ### Added
162
+ - **Mode-aware context packaging.** A `[startup.<mode>]` matrix picks which sources
163
+ compose (persona · profile · learned · company · rules · workflow base/steps) for
164
+ each mode (default / workflow / authoring), with typed per-source compression
165
+ (`[compress.prompts]`).
166
+ - **Personas.** Selectable tone with a universal floor (`gmlw persona list`), a free
167
+ local host greeting at launch, and a first-run persona choice — configured under
168
+ `[companion]`.
169
+ - **Learned notebook.** A portable, user-owned notebook (`profile/me/learned.md`) the
170
+ client mirrors into, read into every client so what one learns they all inherit;
171
+ negatives are first-class.
172
+ - **Rule format.** A domain-neutral `Rule / When / Signals / Strength / Origin`
173
+ (+ optional `Precedence`), captured as a draft during a workflow.
174
+ - **Formal plugins folder.** Reference a caller by id (`~/.gmlw/plugins/<id>/` with a
175
+ `plugin.toml`); `gmlw plugins list`.
176
+ - **Cursor allowance block.** The status line renders cursor's plan pools from an
177
+ optional local cache when the client does not pipe them.
178
+ - **First-run init.** Detect installed clients and seed a filled config with a default.
179
+ - **JSON output** for the listing/reporting commands (`--json`).
180
+ - **Ergonomics.** Client and working-directory preflight, implicit `gmlw <job>`
181
+ (shorthand for `start`), a friendly no-job message, and auto-help for an incomplete
182
+ sub-command.
183
+
184
+ ### Changed
185
+ - Documentation overhaul: task-oriented guides (`docs/USER_GUIDE.md`, `docs/CLI.md`,
186
+ `docs/CONFIGURATION.md`, `docs/CLIENTS.md`, `docs/WORKFLOWS.md`,
187
+ `docs/TROUBLESHOOTING.md`), a client capability matrix in the README, and `DESIGN.md`
188
+ synchronised with the code; `GOVERNANCE.md` corrected. Supported Python stated as
189
+ 3.11–3.14 to match CI.
190
+
191
+ ## [0.1.0] - 2026-07-14
192
+
193
+ First public release — a metering wrapper around ML coding CLIs.
194
+
195
+ ### Added
196
+ - **Jobs & sessions.** Enter at a **job** you tag; the wrapper mints a named,
197
+ resumable **session** on the client and persists it. `gmlw start`, `jobs`,
198
+ `sessions`, `export`.
199
+ - **Four clients** driven the same way: **claude**, **cursor**, **codex**, **vibe**.
200
+ Which one is config-driven (`[client]`) or `--client`.
201
+ - **Metering relay.** A local, capability-URL-authenticated relay records **per-turn
202
+ tokens and cost** for the metered clients (claude/codex/vibe) into a SQLite ledger;
203
+ `gmlw export` reports per-turn rows, per-model totals, and per-session cost.
204
+ - **Status line** for the clients that host one (claude, cursor): git · folder ·
205
+ model · context% · a client-specific allowance block, plus a per-session and
206
+ per-job usage footer.
207
+ - **Workflows.** Author a small operating context once (`gmlw workflow new`, a warm
208
+ create-workflow interview) and launch a job with it; context is compiled from a
209
+ shared base + your profile + rules + the workflow steps, through an interceptor
210
+ chain (with opt-in context compression via `generic-ml-cache`).
211
+ - **Durable provenance.** The exact compiled context is written per session
212
+ (`contexts/<job>/<session>.context.md`); an **opt-in transcript** keeps each metered
213
+ call's request/response/usage.
214
+ - **Credentials.** Per-workflow secrets stored `0600` and injected into the client's
215
+ environment at launch (`gmlw creds set`).
216
+ - **Storage.** A single SQLite ledger (`~/.gmlw/ledger.db`, WAL) for jobs, sessions,
217
+ per-turn usage, and session costs.
218
+ - **Safety.** Validated `JobId` + filesystem containment under an owner-only `~/.gmlw`;
219
+ never overwriting an unparseable client-settings or credentials file.
220
+
221
+ ### Engineering
222
+ - Hexagonal (ports & adapters), enforced by `import-linter`; strict `ruff` + `pyright`
223
+ over `src` and `tests`; `nox` gates mirrored by CI across Python 3.11–3.14; a
224
+ server-side no-AI-attribution check and branch protection.
225
+
226
+ [Unreleased]: https://github.com/danielslobozian/generic-ml-wrapper/compare/v0.5.0...HEAD
227
+ [0.5.0]: https://github.com/danielslobozian/generic-ml-wrapper/compare/v0.4.0...v0.5.0
228
+ [0.4.0]: https://github.com/danielslobozian/generic-ml-wrapper/compare/v0.3.0...v0.4.0
229
+ [0.3.0]: https://github.com/danielslobozian/generic-ml-wrapper/compare/v0.2.0...v0.3.0
230
+ [0.2.0]: https://github.com/danielslobozian/generic-ml-wrapper/compare/v0.1.0...v0.2.0
231
+ [0.1.0]: https://github.com/danielslobozian/generic-ml-wrapper/releases/tag/v0.1.0
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.4
2
2
  Name: generic-ml-wrapper
3
- Version: 0.3.0
3
+ Version: 0.5.0
4
4
  Summary: A wrapper around an ML coding CLI: enter at a job, get a named resumable session, optionally driven by a workflow.
5
5
  Project-URL: Homepage, https://github.com/danielslobozian/generic-ml-wrapper
6
6
  Project-URL: Repository, https://github.com/danielslobozian/generic-ml-wrapper
@@ -26,6 +26,8 @@ Requires-Python: >=3.11
26
26
  Requires-Dist: generic-ml-cache-adapters>=1.0.0
27
27
  Requires-Dist: generic-ml-cache-bootstrap>=1.0.0
28
28
  Requires-Dist: generic-ml-cache-core>=1.0.0
29
+ Requires-Dist: pydantic-settings>=2
30
+ Requires-Dist: tomlkit>=0.13
29
31
  Provides-Extra: dev
30
32
  Requires-Dist: import-linter>=2; extra == 'dev'
31
33
  Requires-Dist: pyright==1.1.411; extra == 'dev'
@@ -0,0 +1,280 @@
1
+ <!-- SPDX-FileCopyrightText: 2026 Daniel Slobozian -->
2
+ <!-- SPDX-License-Identifier: Apache-2.0 -->
3
+
4
+ # Roadmap
5
+
6
+ Where `generic-ml-wrapper` has been and where it's going. Direction, not a
7
+ commitment — see [CHANGELOG.md](CHANGELOG.md) for what has actually shipped.
8
+
9
+ The through-line: turn a metering wrapper into something a person *wants* to open —
10
+ a companion that carries **who you are and how you work across every client**, and
11
+ stays a wrapper (it extends each client; it never reimplements one).
12
+
13
+ ## Shipped
14
+
15
+ ### 0.1.0
16
+ The metering foundation — jobs & sessions, four clients (claude / cursor / codex /
17
+ vibe), a per-turn metering relay into a SQLite ledger, a client-aware status line,
18
+ and the workflow system. On PyPI.
19
+
20
+ ### 0.2.0 — the companion
21
+ Everything a session inherits, and the ergonomics around it.
22
+
23
+ - **First-run init** — detect installed clients, seed a filled config with a default.
24
+ - **Config-driven, mode-aware context packaging** — a `[startup.<mode>]` matrix picks
25
+ which sources compose (persona · profile · learned · company · rules · workflow
26
+ base/steps) for each mode (default / workflow / authoring), with **typed per-source
27
+ compression** (`[compress.prompts]`).
28
+ - **Personas** — selectable tone with a universal floor; `gmlw persona list`; a free,
29
+ local **host greeting** at launch; a first-run persona choice.
30
+ - **Learned** — a portable, user-owned notebook the client mirrors into, read into
31
+ every client so what one learns they all inherit; **negatives are first-class**.
32
+ - **Rule format** — a domain-neutral `Rule / When / Signals / Strength / Origin`
33
+ (+ optional `Precedence`), with in-session capture during a workflow.
34
+ - **Formal plugins folder** — reference a caller by id (`~/.gmlw/plugins/<id>/`),
35
+ `gmlw plugins list`.
36
+ - **Cursor allowance block** — the status line renders cursor's plan pools from an
37
+ optional local cache when the client doesn't pipe them.
38
+ - **Ergonomics** — client & working-directory preflight (clear guidance instead of a
39
+ cryptic crash), implicit `gmlw <job>` (shorthand for `start`), a friendly no-job
40
+ message, and auto-help for an incomplete sub-command.
41
+ - **Create-workflow** — the interviewer assesses each step's *codeability* and offers
42
+ to script the mechanical ones, keeping the AI for the judgment steps.
43
+
44
+ ### 0.3.0 — lifecycle action hooks
45
+ gmlw already hooks **content**: the interceptor chain transforms context sections at
46
+ compile time and wire traffic at relay time (e.g. anonymisation on `request`). It now
47
+ also hooks **actions** — "after this phase, run something" — at two lifecycle seams
48
+ that bracket the client run:
49
+
50
+ - **`pre-launch`** — after the context is compiled and the caller resolved, before the
51
+ client starts. Knows the client, the working directory, and the session. For
52
+ per-client setup: deploying skills/rules into a client's *native* mechanism, writing
53
+ MCP config, warming a cache.
54
+ - **`post-session`** — after the client exits. Knows the exit code and the session.
55
+ For cleanup, notification, archival, or roll-up.
56
+
57
+ Config mirrors the existing extension points — a `[[hooks]]` entry with a `phase`, a
58
+ `spec` (a plugin id or `path.py:Class`), and an optional `client` scope — under the
59
+ same trusted-code boundary as `[[interceptors]]`, `[callers]`, and plugins. Hooks are
60
+ best-effort and never break a launch. Both launch paths (`start`, `workflow new`) route
61
+ through one shared launch sequence, so the seams bracket every run. Shipped with a
62
+ built-in `SessionLogger` reference hook.
63
+
64
+ **Still open — the flagship example hook:** a **cross-client skills/rules deployer** that
65
+ consumes a git repo of skills and installs them, per client, as faithfully as each
66
+ client's format allows. It is its own initiative (per-client format knowledge), built on
67
+ the seam above; the infrastructure is done and waiting for it.
68
+
69
+ #### Rule lifecycle
70
+ 0.2.0 shipped rule *capture*, but only inside a workflow (the directive lived in the
71
+ workflow base) and without dedup or a path to code. The rule loop is now rounded out —
72
+ the directive moved out of the workflow base to the head of the always-on `rules`
73
+ context source (verbatim, gmlw's voice), and the source is active by default in a plain
74
+ start:
75
+
76
+ - **Rule proposal in normal usage** — the "offer to record a rule" directive is
77
+ always-on, so a demanded correction becomes a draft rule wherever it happens, not only
78
+ inside a workflow.
79
+ - **Existing-rule check (dedup / update-not-duplicate)** — before proposing, the client
80
+ reads the existing rules and updates or supersedes a matching one instead of stacking a
81
+ near-duplicate, mirroring the learned notebook's supersede-on-contradiction.
82
+ - **Rule → code feasibility** — a captured rule is judged for *mechanical
83
+ enforceability* and, if so, offered as a script/check rather than a reminder — the
84
+ step-codeability logic (create-workflow) generalised from workflow steps to rules, and
85
+ a natural future `pre-launch` hook consumer.
86
+
87
+ ### 0.4.0 — first run: the forced setup that shapes every session
88
+ One job, done properly: a **mandatory `init`** every user passes through once — new or
89
+ existing — because it establishes the model the rest of the app runs on and migrates the
90
+ on-disk layout to match. It is a *forced update*: the old single-context folder layout is
91
+ wrapped into the new one non-destructively, and role + environment are chosen before the
92
+ first session. The discoverability work that used to sit here has moved to 0.5.0; this
93
+ release is the init and the two concepts under it.
94
+
95
+ **The movie-set model.** A launch composes four axes — three describe *you and your
96
+ context* (they become the briefing the wrapper hands the companion), one describes *the
97
+ companion itself*:
98
+
99
+ - **Me — the actor.** Who you are, invariant across everything: name, language, the
100
+ journal. Spans every launch.
101
+ - **Role — the character.** The functional hat you're wearing (software engineer, product
102
+ owner, QA, a private individual buying groceries). A *lens over `me`*, not a copy — it
103
+ parameterises the `me`-extraction (which facets are in scope) and scopes rules/learned.
104
+ You stay one person; you play different parts.
105
+ - **Environment — the movie set.** Where the work happens (work-at-a-company, a personal
106
+ project, open source). It swaps the *place-specific context* — today's forced `company`
107
+ becomes one environment's bucket. Changing environment changes the set and the
108
+ guidelines, not who you are.
109
+ - **Persona — the director.** The one entity you actually talk to: the companion. Persona
110
+ is *its* manner (manner over method — see 0.6.0).
111
+
112
+ gmlw itself is the **stage organiser / assistant director**: it builds the set, arranges
113
+ costume and makeup, briefs the director, then cedes the screen ([[wrapper-not-standalone]]).
114
+
115
+ - **Forced init (`gmlw init`, and the gate)** — init is both a command and a first-class
116
+ gate: a stored marker records that it ran, and bare `gmlw` routes an un-initialised or
117
+ old-layout install into it before anything else runs. It captures, in order, **language
118
+ → name → role → environment → persona**, then does the technical client step. Each
119
+ answer carries a sensible default so the forced pass stays short.
120
+ - **Two new concepts — role & environment** — the two axes above, made real: `default_role`
121
+ and `default_environment` in `config.toml` (read directly — no registry yet), a
122
+ `profile/roles/<role>/` folder for role-scoped rules/learned, and an
123
+ `environments/<env>/` folder for place-specific context. Both are changeable later (via
124
+ the `config` commands landing in 0.5.0). This **resolves the parked _profiles_ fork** —
125
+ `me` spans, role is a lens over it, environment is the external place — by splitting one
126
+ container into two orthogonal axes.
127
+ - **Forced migration** — the existing global layout (`profile/company`, single-context
128
+ `rules`) is wrapped non-destructively into the new shape — `profile/company` →
129
+ `environments/work/`, a `default` role seeded — surfacing exactly what moved
130
+ ([[no-silent-removals]]). No install is left on the old layout.
131
+ - **Localised setup (EN / FR)** — onboarding strings move into a keyed language file
132
+ (`i18n/en.*`, `i18n/fr.*`) read through a small `t(key, lang)` lookup with English
133
+ fallback; the default is seeded from `$LANG`. Mechanism built to extend; content scoped
134
+ to onboarding for now.
135
+ - **`language` — our voice only** — a `language` setting fixes the language gmlw speaks to
136
+ *you* (onboarding now, receipts/help later). It does **not** force the companion's
137
+ language: only Claude Code exposes a real language setting; Cursor, Codex and Vibe have
138
+ none, so pushing it would be a leaky per-client hack for one client out of four. gmlw
139
+ speaks your language; the companion stays as its own config leaves it.
140
+ - **A working client — the one hard requirement** — after the human setup, the technical
141
+ step. *Fast path*: detect an installed, authenticated client and reach a session in
142
+ seconds. *Guided path* when none is found: a subscription→client map (*"do you pay for
143
+ Claude / ChatGPT / Cursor / Mistral?"*), the **official install command for the detected
144
+ OS** (Windows / Linux / macOS), and **guide-and-verify auth** — print the exact login
145
+ command, poll readiness until it goes green (guide, don't drive).
146
+ - **A default persona** — the experience-defining choice, offered with a one-line
147
+ description per persona so it can be made without reading docs.
148
+
149
+ Deferred but homed: a detached model call to synthesise a `role.md` for an unfamiliar role
150
+ (e.g. product owner) is a natural future consumer of the 0.3.0 `pre-launch` seam — parked,
151
+ not built. **Persona *previews* moved to 0.7.0** — sample lines are only worth shipping once
152
+ the personas behind them are proven to actually differ, and doing them in French requires
153
+ localising persona content, which is the same job (see 0.7.0).
154
+
155
+ ### 0.5.0 — discoverability & progressive disclosure
156
+ Reframed around the real metric for a new user: **time to a first live session**, and then
157
+ revealing depth *gradually*. Because the wrapper cedes the screen to the client, discovery
158
+ lives in the thin surfaces **around** a run — bare-command indexes, the return, and ambient
159
+ context pushed *into* the client — never a persistent UI over a live session.
160
+
161
+ - **App-wide localisation** (`0fe8c0a`) — the init-only JSON catalogue now spans the whole
162
+ app: every user-facing message **and every diagnostic log line** renders through a
163
+ process-global active localiser (`i18n.set_active`/`active`/`t`), English-fallback then
164
+ raw-key safe. A catalogue-drift guard keeps EN/FR key sets identical. Localising the logs
165
+ too was a deliberate choice, not the usual English-only-logs convention.
166
+ - **Config registry** (`9fe05b6`) — a `pydantic-settings` model is the typed source of
167
+ truth for every settable scalar key (type, default, allowed values, description);
168
+ `registry_rows()`/`coerce()`/`load()` drive help, `config`, and validation. `config.py`
169
+ sources its defaults from it (no duplicated literals) while keeping its tolerant reads.
170
+ *Scope:* scalar keys only — the structural matrices (`[[hooks]]`, `[[interceptors]]`,
171
+ `[startup.*.context]`, `[compress.prompts]`) stay hand-rolled, a deferred follow-up.
172
+ - **`config` commands** (`0db3ca7`) — `config list / get / set`, rendering + validating
173
+ against the registry; `set` merges through the shared tomlkit writer (comments preserved,
174
+ never rewritten) and surfaces the old→new change. The home for changing `default_role` /
175
+ `default_environment` after init.
176
+ - **Bare `gmlw` + `gmlw help`** (`61bb15f`) — bare `gmlw` is first-run-aware: fresh install
177
+ → init; thereafter → a grouped capability index (*launch / inspect / author*) with a
178
+ next-action footer. `gmlw help <topic>` explains the core concepts (job-vs-workflow,
179
+ start-vs-run, personas, cost). `--help` keeps the argparse view.
180
+ - **Greeting → context** (`85a77da`) — the launch-time host greeting (structurally
181
+ invisible once the client clears the screen) is retired from stderr and injected into a
182
+ new session's context, so the client renders it in-band. `_farewell` (exit) stays.
183
+ - **Exit receipt + ambient card** (`f36fcd8`) — on the return, a persistent receipt: this
184
+ session's and the job's cost, the resume/report commands, and one usage-driven,
185
+ suppressible tip (shown once each; `[hints] show` to disable). `StartJob` returns a
186
+ `StartJobResult` so the receipt can name the session. The off-by-default ambient
187
+ capability card (`[ambient] capability_card`) injects a "how do I …" gmlw card into the
188
+ context. **Authoring-cost visibility deferred** — the authoring bucket is deliberately
189
+ kept out of `gmlw jobs`; surfacing its spend cleanly is its own design.
190
+ - **`workflow edit`** (`0445c6d`) — amend an existing workflow in an authoring session
191
+ (opens its folder, never creates/overwrites; unknown name → clean error).
192
+ - **Already shipped earlier, verified here:** **`--version`** (surfaces the running version)
193
+ and **robustness** (clean Ctrl+C / SIGTERM interrupt-exit) were already present; their few
194
+ strings were folded into the app-wide localisation pass.
195
+
196
+ ## Planned
197
+
198
+ ### 0.6.0 — the workflow, first-class
199
+ Two ways people relate to a workflow: *applied to a job* (a ticket it treats), or *the
200
+ recurring job itself* (a repeatable extraction). Make both first-class, and make
201
+ authoring one a guided conversation — because most people don't arrive with a clear
202
+ model of their own process.
203
+
204
+ - **`run <workflow>`** — launch a workflow directly; the job defaults to the workflow
205
+ name and sessions accumulate under it. `start <job> [--workflow]` stays for the
206
+ applied case. A trigger-gated pre-launch chooser fills only *missing* arguments and
207
+ always echoes the equivalent one-liner, so interactive use teaches the fast path
208
+ (full argv never gains a prompt).
209
+ - **Name at the end, not the start** — `workflow new` no longer demands a name up front;
210
+ forcing one presumes the user already knows the shape. The authoring conversation
211
+ shapes the workflow and *proposes* the name at convergence. (Distinct from
212
+ `run <workflow>`, where an existing, recurring workflow's name has already earned it.)
213
+ - **Facilitative + constructive authoring** — the create-workflow conversation as a
214
+ blended consultant. Two axes: *facilitate* (a **parking lot** for tangents so nothing
215
+ is lost, reflective listening, a **diverge → converge** phase model, and **process
216
+ leveling** to answer "step or its own workflow?") and *contribute* (start in inquiry
217
+ and **move to expert when warranted**, propose the upstream/downstream stages the
218
+ author omitted, surface **implications** they haven't hit) — bounded by guardrails
219
+ against railroading a novice and a **consent gate** on anything personal. Authoring
220
+ keeps distilled state (a draft plus the parking lot) as files in the workflow folder,
221
+ so it survives context compaction.
222
+ - **Per-workflow persona, composed as voice over method** — a workflow can carry its own
223
+ persona instead of the single global tone. Persona is **manner**; the authoring
224
+ behavior above is **method**. They compose by layer: the method lives in the mode
225
+ floor (guaranteed, persona-independent), and the persona colors delivery on top — it
226
+ can voice the invariants but not break them. Personas carry an **authoring-friendliness**
227
+ expectation, so a voice built around brutal efficiency can't sabotage the facilitation.
228
+
229
+ ### 0.7.0 — personas, proven (and multilingual)
230
+ Today "personas shape tone" is an **untested claim**. A persona ships a tone block, but
231
+ nothing demonstrates that `mentor` and `terse` actually answer differently — and the tone
232
+ block is injected *on top of* each client's own system prompt, which may simply swamp it.
233
+ This release makes the claim falsifiable first, then builds on it. Its own initiative: it
234
+ is an evaluation loop, not a feature.
235
+
236
+ - **A persona evaluation harness** — the core of the release. Run every persona against a
237
+ fixed question set and compare. Two design calls, both deliberate:
238
+ - **Questions curated against the declared dimensions, not scraped.** Generic benchmark
239
+ sets measure *correctness* and would show almost no tone variance. Each persona already
240
+ declares `Warmth · Verbosity · Formality · Proactivity`, so the set probes exactly those,
241
+ plus the high-divergence moments — being corrected, delivering bad news, being handed a
242
+ half-formed request.
243
+ - **A distinctness metric, not eyeballing.** A judge scores each answer on the declared
244
+ dimensions and yields a **pairwise distance matrix**, so collapsed pairs are named rather
245
+ than sensed (`plain`/`terse` and `companion`/`mentor` are the likely offenders). Runs go
246
+ through **generic-ml-cache**, so re-running one persona after a tweak replays the other
247
+ four for free and keeps iterations deterministic and diffable.
248
+ - **Tune until distinct** — the back-and-forth the harness exists to serve: adapt each tone
249
+ block, re-run, watch the matrix separate. The honest possible outcome is that personas must
250
+ become **bolder** to survive a client's own voice — or that they differentiate on some
251
+ clients and not others. That finding is a deliverable, not a failure.
252
+ - **Persona content becomes multilingual** — required for previews in French, and a real gap
253
+ today: only the *prompt header* is localised, so a French user gets an English description
254
+ and an **English greeting at every launch**. This is not a strings port: French carries a
255
+ `tu`/`vous` register English cannot express, and a butler is *defined* by register — so the
256
+ directives themselves need per-language authoring, not translation.
257
+ - Layout: a **folder per persona** — a language-neutral `body.md` (dimensions, do/don't)
258
+ plus `en.md` / `fr.md` carrying description, greeting, register notes, and samples.
259
+ - **A flat `<name>.md` must keep working.** Users author personas by dropping a file in
260
+ `~/.gmlw/personas/`; a folder gets the full treatment, a flat file behaves as it does
261
+ today (English, no samples, falls back to its description). Breaking user-authored
262
+ personas to gain previews would be a bad trade.
263
+ - **Previews, as a byproduct** — once the answers are generated, reviewed, and committed as
264
+ data, the first-run chooser can show a real sample line per persona. It stays **static** at
265
+ runtime: generated at authoring time, read from disk, **zero tokens and zero latency**, so
266
+ choosing still costs nothing. Length forces a split — one short line inline in the chooser,
267
+ the full side-by-side behind `gmlw persona preview`. Samples are labelled as indicative:
268
+ they come from one model at authoring time, and the user's client may differ.
269
+
270
+ **Sequencing risk, stated plainly:** 0.6.0 layers *per-workflow persona* on top of personas
271
+ whose distinctness this release is what proves. If the matrix shows they collapse, 0.6.0's
272
+ persona composition needs revisiting — so pull this earlier if that work starts leaning hard
273
+ on the foundation.
274
+
275
+ ## Parked
276
+
277
+ - **External source connectors** — let a workflow pull from external systems (APIs,
278
+ cloud storage, platforms) rather than manual input. Large; its own initiative.
279
+ - Relay extraction to a standalone project (shelved — no second consumer yet).
280
+ - Workspace layout feature.
@@ -1,6 +1,6 @@
1
1
  # Security Policy
2
2
 
3
- `generic-ml-wrapper` is early software (`0.2.0`). It is published
3
+ `generic-ml-wrapper` is early software (`0.5.0`). It is published
4
4
  openly and takes security seriously regardless of maturity. Thank you for helping keep it
5
5
  and its users safe.
6
6
 
@@ -0,0 +1 @@
1
+ 0.5.0