better-code-review-graph 3.29.8__tar.gz → 3.30.0__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 (252) hide show
  1. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/.claude-plugin/plugin.json +9 -54
  2. better_code_review_graph-3.30.0/.gitattributes +7 -0
  3. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/.github/workflows/cd.yml +4 -4
  4. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/.github/workflows/ci.yml +1 -1
  5. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/.gitignore +8 -0
  6. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/AGENTS.md +18 -34
  7. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/CHANGELOG.md +53 -0
  8. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/CLAUDE.md +17 -33
  9. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/Dockerfile +12 -4
  10. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/PKG-INFO +109 -30
  11. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/README.md +108 -29
  12. better_code_review_graph-3.30.0/docker-compose.yml +39 -0
  13. better_code_review_graph-3.30.0/docker-config/config.example.toml +51 -0
  14. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/pyproject.toml +1 -1
  15. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/server.json +6 -30
  16. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/src/crg/cli.py +7 -1
  17. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/src/crg/config.py +10 -0
  18. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/src/crg/docs/graph.md +5 -4
  19. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/src/crg/server.py +18 -3
  20. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/src/crg/tools.py +3 -3
  21. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_embeddings.py +4 -14
  22. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_g6_ux_status_accuracy.py +1 -4
  23. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_server.py +81 -0
  24. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_setup_tool.py +0 -1
  25. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_summarize_graph_dispatch.py +6 -11
  26. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_summarizer.py +1 -1
  27. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_v17_coverage_gaps.py +0 -3
  28. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/uv.lock +7 -7
  29. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/.agents/skills/build-graph/SKILL.md +0 -0
  30. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/.agents/skills/review-delta/SKILL.md +0 -0
  31. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/.agents/skills/review-pr/SKILL.md +0 -0
  32. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/.coderabbit.yaml +0 -0
  33. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/.dockerignore +0 -0
  34. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/.github/CODEOWNERS +0 -0
  35. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/.github/ISSUE_TEMPLATE/bug_report.md +0 -0
  36. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/.github/ISSUE_TEMPLATE/feature_request.md +0 -0
  37. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/.github/PULL_REQUEST_TEMPLATE.md +0 -0
  38. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/.github/review-learnings.md +0 -0
  39. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/.github/rulesets/main.json +0 -0
  40. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/.infisical.json +0 -0
  41. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/.jules/bolt.md +0 -0
  42. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/.jules/sentinel.md +0 -0
  43. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/.pre-commit-config.yaml +0 -0
  44. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/BREAKING_CHANGES.md +0 -0
  45. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/CODE_OF_CONDUCT.md +0 -0
  46. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/CONTRIBUTING.md +0 -0
  47. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/LICENSE +0 -0
  48. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/LICENSE-MIT +0 -0
  49. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/NOTICE +0 -0
  50. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/SECURITY.md +0 -0
  51. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/alembic.ini +0 -0
  52. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/docs/HANDOVER.md +0 -0
  53. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/hooks/_graph_state.py +0 -0
  54. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/hooks/branch-sync-check.py +0 -0
  55. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/hooks/hooks.json +0 -0
  56. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/hooks/stale-graph-check.py +0 -0
  57. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/migrations/README +0 -0
  58. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/migrations/env.py +0 -0
  59. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/migrations/script.py.mako +0 -0
  60. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/migrations/versions/001_baseline.py +0 -0
  61. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/migrations/versions/002_summary_columns.py +0 -0
  62. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/migrations/versions/003_federation.py +0 -0
  63. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/migrations/versions/004_security_tags.py +0 -0
  64. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/migrations/versions/005_temporal_columns.py +0 -0
  65. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/migrations/versions/006_commits_table.py +0 -0
  66. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/renovate.json +0 -0
  67. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/rules/heuristic/cwe-22-path-traversal.yaml +0 -0
  68. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/rules/heuristic/cwe-78-shell-command-sink.yaml +0 -0
  69. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/rules/heuristic/cwe-89-sql-string-format.yaml +0 -0
  70. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/rules/heuristic/cwe-95-eval-on-input.yaml +0 -0
  71. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/rules/heuristic/hardcoded-secret.yaml +0 -0
  72. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/rules/semgrep/curated.yaml +0 -0
  73. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/scripts/benchmark_row_deserialization.py +0 -0
  74. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/scripts/enforce-commit.sh +0 -0
  75. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/scripts/preserve-diacritics.py +0 -0
  76. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/scripts/test_preserve_diacritics.py +0 -0
  77. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/semantic-release.toml +0 -0
  78. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/skills/impact-audit/SKILL.md +0 -0
  79. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/skills/onboard-repo/SKILL.md +0 -0
  80. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/skills/refactor-check/SKILL.md +0 -0
  81. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/skills/review-delta/SKILL.md +0 -0
  82. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/skills/review-pr/SKILL.md +0 -0
  83. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/skills/security-sweep/SKILL.md +0 -0
  84. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/smithery.yaml +0 -0
  85. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/src/crg/__init__.py +0 -0
  86. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/src/crg/__main__.py +0 -0
  87. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/src/crg/credential_state.py +0 -0
  88. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/src/crg/docs/__init__.py +0 -0
  89. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/src/crg/docs/config.md +0 -0
  90. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/src/crg/docs/query.md +0 -0
  91. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/src/crg/docs/recipes.md +0 -0
  92. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/src/crg/docs/review.md +0 -0
  93. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/src/crg/docs/security.md +0 -0
  94. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/src/crg/embeddings.py +0 -0
  95. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/src/crg/exporter.py +0 -0
  96. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/src/crg/federation.py +0 -0
  97. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/src/crg/graph.py +0 -0
  98. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/src/crg/importer.py +0 -0
  99. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/src/crg/incremental.py +0 -0
  100. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/src/crg/parser.py +0 -0
  101. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/src/crg/py.typed +0 -0
  102. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/src/crg/reranker.py +0 -0
  103. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/src/crg/resolver/__init__.py +0 -0
  104. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/src/crg/resolver/_types.py +0 -0
  105. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/src/crg/resolver/fallback.py +0 -0
  106. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/src/crg/resolver/go.py +0 -0
  107. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/src/crg/resolver/java.py +0 -0
  108. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/src/crg/resolver/python.py +0 -0
  109. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/src/crg/resolver/rust.py +0 -0
  110. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/src/crg/resolver/typescript.py +0 -0
  111. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/src/crg/security/__init__.py +0 -0
  112. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/src/crg/security/heuristic.py +0 -0
  113. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/src/crg/security/semgrep_engine.py +0 -0
  114. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/src/crg/summarizer.py +0 -0
  115. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/src/crg/temporal.py +0 -0
  116. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/src/crg/xpia.py +0 -0
  117. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/__init__.py +0 -0
  118. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/conftest.py +0 -0
  119. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/conftest_e2e.py +0 -0
  120. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/fixtures/Sample.cs +0 -0
  121. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/fixtures/SampleJava.java +0 -0
  122. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/fixtures/callback_refs.js +0 -0
  123. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/fixtures/caller_example.py +0 -0
  124. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/fixtures/multi_call_example.py +0 -0
  125. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/fixtures/sample.c +0 -0
  126. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/fixtures/sample.cpp +0 -0
  127. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/fixtures/sample.kt +0 -0
  128. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/fixtures/sample.php +0 -0
  129. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/fixtures/sample.rb +0 -0
  130. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/fixtures/sample.sol +0 -0
  131. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/fixtures/sample.swift +0 -0
  132. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/fixtures/sample_go.go +0 -0
  133. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/fixtures/sample_python.py +0 -0
  134. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/fixtures/sample_rust.rs +0 -0
  135. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/fixtures/sample_typescript.ts +0 -0
  136. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/fixtures/test_sample.py +0 -0
  137. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/integration/__init__.py +0 -0
  138. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/integration/test_multi_user_remote.py +0 -0
  139. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_alembic_baseline.py +0 -0
  140. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_as_of_diff.py +0 -0
  141. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_auto_resolve_ambiguity.py +0 -0
  142. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_bare_suffix_correctness.py +0 -0
  143. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_callers_not_found_reason.py +0 -0
  144. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_callers_of_coverage.py +0 -0
  145. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_cli.py +0 -0
  146. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_cli_contracts.py +0 -0
  147. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_cli_control_plane.py +0 -0
  148. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_cohere_dimensions.py +0 -0
  149. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_commits_migration.py +0 -0
  150. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_coverage_gaps.py +0 -0
  151. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_credential_state.py +0 -0
  152. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_credential_state_dehosted.py +0 -0
  153. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_crg_protocol_harness.py +0 -0
  154. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_custom_local_model.py +0 -0
  155. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_dynamic_dispatch_hints.py +0 -0
  156. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_e2e.py +0 -0
  157. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_embedder_retry_classification.py +0 -0
  158. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_embedding_migration.py +0 -0
  159. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_embeddings_extra.py +0 -0
  160. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_export_graph_dispatch.py +0 -0
  161. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_export_graph_dispatch_path_traversal.py +0 -0
  162. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_exporter.py +0 -0
  163. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_federation.py +0 -0
  164. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_federation_migration.py +0 -0
  165. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_federation_smoke.py +0 -0
  166. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_full_live.py +0 -0
  167. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_get_docs_section_error_path.py +0 -0
  168. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_get_review_context_no_diff.py +0 -0
  169. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_graph.py +0 -0
  170. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_graph_batch_fetch.py +0 -0
  171. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_graph_extra.py +0 -0
  172. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_graph_get_edges_by_targets_coverage.py +0 -0
  173. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_graph_stats_error_coverage.py +0 -0
  174. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_handle_not_found_coverage.py +0 -0
  175. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_help_recipes.py +0 -0
  176. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_heuristic.py +0 -0
  177. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_home_isolation.py +0 -0
  178. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_http_mcp_e2e.py +0 -0
  179. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_hull_cells.py +0 -0
  180. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_impact_payload_truncation.py +0 -0
  181. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_implements_edges.py +0 -0
  182. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_import_graph_dispatch.py +0 -0
  183. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_importer.py +0 -0
  184. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_incremental.py +0 -0
  185. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_incremental_extra.py +0 -0
  186. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_kind_filter_optimization.py +0 -0
  187. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_kinds_error_path.py +0 -0
  188. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_list_graph_stats_coverage.py +0 -0
  189. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_list_kinds_coverage.py +0 -0
  190. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_live_mcp.py +0 -0
  191. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_local_reranker.py +0 -0
  192. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_lookup_node_directly.py +0 -0
  193. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_multi_user.py +0 -0
  194. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_multilang.py +0 -0
  195. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_no_stale_package_name.py +0 -0
  196. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_oci_publication_sunset.py +0 -0
  197. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_oserror_coverage.py +0 -0
  198. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_parser.py +0 -0
  199. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_parser_extra.py +0 -0
  200. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_parser_grammar_failure.py +0 -0
  201. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_parser_source_text.py +0 -0
  202. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_path_resolution_fallback.py +0 -0
  203. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_payload_estimation.py +0 -0
  204. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_per_sub_graph_db.py +0 -0
  205. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_performance_n1.py +0 -0
  206. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_performance_n1_embeddings.py +0 -0
  207. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_phase_3_smoke.py +0 -0
  208. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_plugin_assets.py +0 -0
  209. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_pre_2_0_backup.py +0 -0
  210. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_renamed_in_diff.py +0 -0
  211. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_renamed_in_diff_exception.py +0 -0
  212. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_repo_filter.py +0 -0
  213. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_repo_filter_batching.py +0 -0
  214. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_repo_id_persistence.py +0 -0
  215. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_resolve_search_candidates.py +0 -0
  216. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_resolver_dispatcher.py +0 -0
  217. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_resolver_fallback.py +0 -0
  218. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_resolver_go.py +0 -0
  219. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_resolver_java.py +0 -0
  220. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_resolver_python.py +0 -0
  221. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_resolver_rust.py +0 -0
  222. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_resolver_typescript.py +0 -0
  223. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_response_header.py +0 -0
  224. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_response_payload_honesty.py +0 -0
  225. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_review_delta_line_shifts.py +0 -0
  226. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_review_languages_filter.py +0 -0
  227. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_reviewer_summary.py +0 -0
  228. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_search_edges_by_target_names.py +0 -0
  229. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_search_keyword_warning.py +0 -0
  230. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_security_injection.py +0 -0
  231. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_security_path_traversal.py +0 -0
  232. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_security_scan_batching.py +0 -0
  233. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_security_tags_migration.py +0 -0
  234. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_security_tool.py +0 -0
  235. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_semgrep_engine.py +0 -0
  236. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_server_help_coverage.py +0 -0
  237. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_serverinfo_version.py +0 -0
  238. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_shares_state_edges.py +0 -0
  239. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_spot_check.py +0 -0
  240. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_sql_security.py +0 -0
  241. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_stdio_direct.py +0 -0
  242. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_summary_schema.py +0 -0
  243. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_temporal.py +0 -0
  244. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_temporal_migration.py +0 -0
  245. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_tested_by_edges.py +0 -0
  246. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_tools.py +0 -0
  247. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_tools_build_error_coverage.py +0 -0
  248. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_tools_full.py +0 -0
  249. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_tools_get_docs_section_coverage.py +0 -0
  250. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_typescript_coverage_gaps.py +0 -0
  251. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_xpia.py +0 -0
  252. {better_code_review_graph-3.29.8 → better_code_review_graph-3.30.0}/tests/test_xpia_adversarial.py +0 -0
@@ -1,7 +1,7 @@
1
1
  {
2
2
  "name": "better-code-review-graph",
3
3
  "description": "Knowledge graph for token-efficient code reviews -- semantic search and call-graph resolution across your codebase.",
4
- "version": "3.29.8",
4
+ "version": "3.30.0",
5
5
  "userConfig": {
6
6
  "EMBEDDING_MODELS": {
7
7
  "type": "string",
@@ -45,56 +45,17 @@
45
45
  "description": "Apply explicit L2 normalization for an external ID without a manifest.",
46
46
  "required": false
47
47
  },
48
- "SUMMARY_MODELS": {
48
+ "HULL_EMBED_API_KEY": {
49
49
  "type": "string",
50
- "title": "Summarizer model chain (optional)",
51
- "description": "CSV 'provider/model,...' for graph summarize; selects the first entry without runtime fallback. Empty = summaries disabled. Example: openrouter/minimax/minimax-m3:free",
52
- "required": false
53
- },
54
- "EMBEDDING_API_BASE": {
55
- "type": "string",
56
- "title": "Embedding endpoint (optional)",
57
- "description": "Provider-compatible endpoint or CF AI Gateway embedding URL; SSRF-guarded.",
58
- "required": false
59
- },
60
- "LLM_API_BASE": {
61
- "type": "string",
62
- "title": "Completion endpoint (optional)",
63
- "description": "Provider-compatible base URL or CF AI Gateway OpenRouter route; SSRF-guarded.",
64
- "required": false
65
- },
66
- "OPENROUTER_API_KEY": {
67
- "type": "string",
68
- "title": "OpenRouter API key (optional)",
69
- "description": "Credential for explicitly selected openrouter/ completion models.",
70
- "sensitive": true,
71
- "required": false
72
- },
73
- "JINA_AI_API_KEY": {
74
- "type": "string",
75
- "title": "Jina AI API key (optional)",
76
- "description": "Credential for explicitly selected jina_ai/ embedding models. An empty EMBEDDING_MODELS selection uses local ONNX; keys alone do not select cloud. https://jina.ai/api-dashboard/",
77
- "sensitive": true,
78
- "required": false
79
- },
80
- "GEMINI_API_KEY": {
81
- "type": "string",
82
- "title": "Gemini API key (optional)",
83
- "description": "Enables gemini/ models referenced in EMBEDDING_MODELS / SUMMARY_MODELS. Free tier available. https://aistudio.google.com/apikey",
84
- "sensitive": true,
85
- "required": false
86
- },
87
- "OPENAI_API_KEY": {
88
- "type": "string",
89
- "title": "OpenAI API key (optional)",
90
- "description": "Enables openai/ models referenced in EMBEDDING_MODELS / SUMMARY_MODELS. https://platform.openai.com/api-keys",
50
+ "title": "Embedding cell API key (optional)",
51
+ "description": "Host-injected key for the [models.embed] cell transport (OpenRouter pre-wired default); overrides the cell api_key. Model name still comes from EMBEDDING_MODELS.",
91
52
  "sensitive": true,
92
53
  "required": false
93
54
  },
94
- "COHERE_API_KEY": {
55
+ "HULL_CHAT_API_KEY": {
95
56
  "type": "string",
96
- "title": "Cohere API key (optional)",
97
- "description": "Enables cohere/ embedding models referenced in EMBEDDING_MODELS. https://dashboard.cohere.com/api-keys",
57
+ "title": "Chat cell API key (optional)",
58
+ "description": "Host-injected key for the [models.chat] cell transport (OpenRouter pre-wired default); enables LLM summaries when the cell model is set.",
98
59
  "sensitive": true,
99
60
  "required": false
100
61
  }
@@ -121,14 +82,8 @@
121
82
  "LOCAL_EMBEDDING_MODEL_FILE": "${user_config.LOCAL_EMBEDDING_MODEL_FILE}",
122
83
  "LOCAL_EMBEDDING_POOLING": "${user_config.LOCAL_EMBEDDING_POOLING}",
123
84
  "LOCAL_EMBEDDING_NORMALIZE": "${user_config.LOCAL_EMBEDDING_NORMALIZE}",
124
- "SUMMARY_MODELS": "${user_config.SUMMARY_MODELS}",
125
- "EMBEDDING_API_BASE": "${user_config.EMBEDDING_API_BASE}",
126
- "LLM_API_BASE": "${user_config.LLM_API_BASE}",
127
- "OPENROUTER_API_KEY": "${user_config.OPENROUTER_API_KEY}",
128
- "JINA_AI_API_KEY": "${user_config.JINA_AI_API_KEY}",
129
- "GEMINI_API_KEY": "${user_config.GEMINI_API_KEY}",
130
- "OPENAI_API_KEY": "${user_config.OPENAI_API_KEY}",
131
- "COHERE_API_KEY": "${user_config.COHERE_API_KEY}"
85
+ "HULL_EMBED_API_KEY": "${user_config.HULL_EMBED_API_KEY}",
86
+ "HULL_CHAT_API_KEY": "${user_config.HULL_CHAT_API_KEY}"
132
87
  }
133
88
  }
134
89
  }
@@ -0,0 +1,7 @@
1
+ # Normalize line endings: the Docker build copies pyproject.toml/uv.lock
2
+ # verbatim and hatchling reads [project.scripts] + force-include from
3
+ # pyproject.toml — CRLF in those files must never reach the image.
4
+ * text=auto
5
+ pyproject.toml text eol=lf
6
+ uv.lock text eol=lf
7
+ Dockerfile text eol=lf
@@ -60,7 +60,7 @@ jobs:
60
60
  # learn the version it WOULD publish, so the PyPI-collision guard below can fire
61
61
  # BEFORE the real run creates a git tag or GitHub Release. Same action + SHA and
62
62
  # same prerelease inputs as the real step, run against the same untouched checkout.
63
- - uses: n24q02m/better-semantic-release@c102741e544c9733760b09d59144f799c51cff8a # v1.7.3
63
+ - uses: n24q02m/better-semantic-release@4bab853128d17c7d8d0173c79883624e77a3ead6 # v1.12.0
64
64
  id: dryrun
65
65
  if: inputs.publish_existing_tag == ''
66
66
  with:
@@ -127,7 +127,7 @@ jobs:
127
127
  echo "::error::Could not verify whether ${PKG}==${COMPUTED_VERSION} exists on PyPI: unexpected HTTP status ${http_code} from ${URL} (likely a transient network/registry error, rate limit, or outage). Refusing to proceed so a genuine collision cannot slip through as a false 'free'. Re-run the release once the registry is reachable."
128
128
  exit 1
129
129
 
130
- - uses: n24q02m/better-semantic-release@c102741e544c9733760b09d59144f799c51cff8a # v1.7.3
130
+ - uses: n24q02m/better-semantic-release@4bab853128d17c7d8d0173c79883624e77a3ead6 # v1.12.0
131
131
  id: release
132
132
  if: inputs.publish_existing_tag == ''
133
133
  with:
@@ -135,7 +135,7 @@ jobs:
135
135
  prerelease: ${{ env.RELEASE_TYPE == 'beta' }}
136
136
  prerelease_token: beta
137
137
 
138
- - uses: python-semantic-release/publish-action@6d89e2de8d40ecb3717316ecbbf15a6fa8ea9039 # v10.6.2
138
+ - uses: python-semantic-release/publish-action@99d4062dd393b5a7a62c5fe9ed78fb364b5674f2 # v10.7.0
139
139
  if: inputs.publish_existing_tag == '' && steps.release.outputs.released == 'true'
140
140
  with:
141
141
  github_token: ${{ steps.app-token.outputs.token }}
@@ -172,7 +172,7 @@ jobs:
172
172
  - uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
173
173
  with:
174
174
  ref: ${{ needs.release.outputs.tag }}
175
- - uses: astral-sh/setup-uv@bec219d24cd3e171d82865faccec33120bb574f4 # v10.1.0
175
+ - uses: astral-sh/setup-uv@c18668ad3cf93ea998bef934396af7bb5c839dc7 # v10.2.0
176
176
  - run: uv build
177
177
  - name: Publish to PyPI
178
178
  run: uv publish
@@ -78,7 +78,7 @@ jobs:
78
78
  uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
79
79
 
80
80
  - name: Install uv
81
- uses: astral-sh/setup-uv@bec219d24cd3e171d82865faccec33120bb574f4 # v10.1.0
81
+ uses: astral-sh/setup-uv@c18668ad3cf93ea998bef934396af7bb5c839dc7 # v10.2.0
82
82
 
83
83
  - name: Setup Python 3.13
84
84
  uses: actions/setup-python@5fda3b95a4ea91299a34e894583c3862153e4b97 # v7.0.0
@@ -95,3 +95,11 @@ clean.ps1
95
95
  *.bak
96
96
  /*.patch
97
97
  /*.diff
98
+
99
+ # local docker deployment config (secrets)
100
+ docker-config/config.toml
101
+
102
+ # e2e scratch (mode-3 isolation runs) — never track
103
+ e2e-mode3-home/
104
+ .e2e_mode3.py
105
+ e2e_token*.txt
@@ -77,23 +77,10 @@ Source files --> Tree-sitter parser --> SQLite graph (nodes + edges)
77
77
  Embedding (cloud backend) + the LLM summarizer dispatch through
78
78
  OpenAI-compatible HTTP clients (`hull_core.providers.openai_spec`).
79
79
 
80
- Per-task model chains, CSV `provider/model,provider/model`, order = fallback. Provider is inferred from the model prefix.
81
-
82
- - `EMBEDDING_MODELS` -- chain embedding. Empty = local ONNX from the fastretrieval built-in registry.
80
+ - `EMBEDDING_MODELS` -- CSV `provider/model,...` selection for the cloud embedding chain. The first entry is active; later entries are retained config, not runtime fallbacks. Empty = local ONNX from the fastretrieval built-in registry.
83
81
  - **Local (default)**: fastretrieval ONNX registry -- zero-config, ~570MB download on first use, 768-dim MRL truncation
84
- - API key follows the `<PROVIDER>_API_KEY` convention. The 7 providers documented below:
85
-
86
- | model prefix | key env var | get it at |
87
- |---|---|---|
88
- | `gemini/` | `GEMINI_API_KEY` | aistudio.google.com/apikey |
89
- | `openai/` (or bare) | `OPENAI_API_KEY` | platform.openai.com |
90
- | `jina_ai/` | `JINA_AI_API_KEY` | jina.ai/api-key |
91
- | `cohere/` | `COHERE_API_KEY` | dashboard.cohere.com |
92
- | `xai/` | `XAI_API_KEY` | console.x.ai |
93
- | `anthropic/` | `ANTHROPIC_API_KEY` | console.anthropic.com |
94
- | `vertex_express/` | `GOOGLE_VERTEX_EXPRESS_API_KEY` | cloud.google.com/vertex-ai/generative-ai/docs/start/express-mode/overview |
95
-
96
- - Custom endpoint (SSRF-guarded): `EMBEDDING_API_BASE` -- custom OpenAI-compatible base URL for cloud embedding (optional)
82
+ - Transport + credentials come from the per-task `[models.<task>]` cell in the instance config (`$CRG_CONFIG_DIR` or `~/.crg/config.toml`), fields `base_url` + `api_key` + `model` — plain OpenAI-spec HTTP through hull-core, OpenRouter pre-wired default (`hull config init`). Env overrides: `HULL_EMBED_API_KEY` / `HULL_CHAT_API_KEY` (host-injected).
83
+ - Model-name prefixes (`cohere/…`, `jina_ai/…`, `gemini/…`, `openrouter/…`) only select wire details (e.g. Cohere `input_type`); they never pick keys or transport. There are no per-vendor API-key env vars — ambient `*_API_KEY` vars are not read.
97
84
  - `DISABLE_LOCAL_EMBED` -- skip the local ONNX download; embedding is `unavailable` unless a cloud chain is configured (`resolve_backend` 3-way: cloud / local / unavailable)
98
85
  - Fixed 768-dim storage keeps the table schema valid across providers. Switching embedding MODEL changes the vector space; embeddings are tagged per provider (`embeddings.provider` column) and `EmbeddingStore.search` restricts the cosine scan to the active provider, so a provider switch just re-embeds rather than mixing incomparable vectors.
99
86
  - Deprecated (honored one release with a warning): singular `EMBEDDING_MODEL` + `EMBEDDING_BACKEND` (backend is now inferred from whether the chain is empty). The old "Jina > Gemini > OpenAI > Cohere" auto-detect router is gone.
@@ -112,28 +99,25 @@ chối, không tự rơi về model mặc định.
112
99
 
113
100
  ### Manual config example
114
101
 
115
- ```json
116
- {
117
- "mcpServers": {
118
- "crg": {
119
- "command": "uvx", "args": ["better-code-review-graph"],
120
- "env": {
121
- "EMBEDDING_MODELS": "jina_ai/jina-embeddings-v5-text-small,gemini/gemini-embedding-001",
122
- "SUMMARY_MODELS": "gemini/gemini-2.5-flash",
123
- "JINA_AI_API_KEY": "jina_xxx",
124
- "GEMINI_API_KEY": "AIza_xxx"
125
- }
126
- }
127
- }
128
- }
102
+ `~/.crg/config.toml`:
103
+
104
+ ```toml
105
+ [models.embed]
106
+ base_url = "https://openrouter.ai/api/v1"
107
+ api_key = "sk-or-..."
108
+ model = "jina-ai/jina-embeddings-v5-text-small"
109
+
110
+ [models.chat]
111
+ base_url = "https://openrouter.ai/api/v1"
112
+ api_key = "sk-or-..."
113
+ model = "minimax/minimax-m3:free"
129
114
  ```
130
115
 
131
116
  ## LLM summarizer (graph `summarize` action)
132
117
 
133
- - `SUMMARY_MODELS` -- ordered summarizer model chain (CSV `provider/model,...`, order = fallback). Empty = summaries disabled. Provider is inferred from the model prefix and must expose a chat-completion API (Jina/Cohere do not).
134
- - Dispatches through OpenAI-compatible chat completions (`hull_core.providers.openai_spec`).
135
- - `LLM_API_BASE` -- custom OpenAI-compatible base URL for the summarizer (SSRF-guarded, optional)
136
- - Deprecated (honored one release with a warning): singular `SUMMARY_MODEL` -- folded into `SUMMARY_MODELS`.
118
+ - `[models.chat]` cell (`base_url` + `api_key` + `model` in `~/.crg/config.toml`, or the `HULL_CHAT_API_KEY` env override) -- an unconfigured cell disables summaries. OpenRouter pre-wired default; the chat model must expose a chat-completion API (embedding-only models do not).
119
+ - Dispatches through OpenAI-compatible chat completions (`hull_core.providers.openai_spec`); the cell `base_url` is SSRF-guarded.
120
+ - The pre-de-host summary-chain, singular-summary-model, and base-URL env vars are removed — dead strings, no longer read.
137
121
 
138
122
  ## Pytest
139
123
 
@@ -2,6 +2,59 @@
2
2
 
3
3
  <!-- version list -->
4
4
 
5
+ ## v3.30.0 (2026-10-06)
6
+
7
+ ### Bug Fixes
8
+
9
+ - **deps**: Lock file maintenance ([#1141](https://github.com/n24q02m/crg/pull/1141),
10
+ [`128bcf6`](https://github.com/n24q02m/crg/commit/128bcf65283f9cc182aaa72ae982b698ba9981cb))
11
+
12
+ - **deps**: Update semgrep to v1.161.0 ([#1140](https://github.com/n24q02m/crg/pull/1140),
13
+ [`afeb04f`](https://github.com/n24q02m/crg/commit/afeb04f46a8f5c56315f9b13fba715954af4ced2))
14
+
15
+
16
+ ## v3.30.0-beta.1 (2026-10-05)
17
+
18
+ ### Bug Fixes
19
+
20
+ - Correct provider docs to per-task cells, drop pre-dehost mode badge
21
+ ([`11d91d7`](https://github.com/n24q02m/crg/commit/11d91d7c4c75588986073c0858ceb976d7e29655))
22
+
23
+ - Ignore e2e_token*.txt scratch files (same rule as mnemo #1287)
24
+ ([`bdc48cf`](https://github.com/n24q02m/crg/commit/bdc48cf157e9c7181531270c525b4e9909a2dc3e))
25
+
26
+ - Purge dead provider env surface in favor of models cells
27
+ ([#1139](https://github.com/n24q02m/crg/pull/1139),
28
+ [`28a046b`](https://github.com/n24q02m/crg/commit/28a046bc2258f4b0b72089f8a698d03e7706fb4c))
29
+
30
+ - Sort imports in tests/test_server.py so CI ruff check passes
31
+ ([`e7d1138`](https://github.com/n24q02m/crg/commit/e7d1138f9925d63918476c76bcc50c76416729f6))
32
+
33
+ - **deps**: Align editable project version in uv.lock to pyproject (uv relock side-effect;
34
+ uv-lock-no-sources hook cannot resolve unpublished hull-core path dep — known de-host state)
35
+ ([`2572c92`](https://github.com/n24q02m/crg/commit/2572c9293e4f144f705a6b17a3f639058e1c0b7f))
36
+
37
+ - **deps**: Update minor dependencies ([#1087](https://github.com/n24q02m/crg/pull/1087),
38
+ [`732ba3c`](https://github.com/n24q02m/crg/commit/732ba3c49e340d9e665acb04fd607bba296dedfd))
39
+
40
+ - **docker**: Drop pyproject-mangling sed; pin LF via .gitattributes
41
+ ([`df1a290`](https://github.com/n24q02m/crg/commit/df1a29097737e3a03ff1a473b963ab8aa32ef47d))
42
+
43
+ - **docker**: Use valid 'server start' subcommand as http ENTRYPOINT
44
+ ([`2ac9b98`](https://github.com/n24q02m/crg/commit/2ac9b98716e907abe07a40c10d4f4e858a8a66fd))
45
+
46
+ - **server**: Enforce no-auth loopback bind on the shared run_http path
47
+ ([`31aeac1`](https://github.com/n24q02m/crg/commit/31aeac1192cf05a4a644960bbf31d1114dbebadc))
48
+
49
+ ### Features
50
+
51
+ - **compose**: Restart unless-stopped for always-on docker topology
52
+ ([`14e87da`](https://github.com/n24q02m/crg/commit/14e87da54271694f30720ce8b099180f2285a947))
53
+
54
+ - **docker**: Self-contained http compose + config example + usage docs
55
+ ([`cba533f`](https://github.com/n24q02m/crg/commit/cba533f43e1a2fcce6ed016cff62b180bb3aa16b))
56
+
57
+
5
58
  ## v3.29.8 (2026-10-05)
6
59
 
7
60
 
@@ -74,28 +74,14 @@ Source files --> Tree-sitter parser --> SQLite graph (nodes + edges)
74
74
  Embedding (cloud backend) + the LLM summarizer dispatch through
75
75
  OpenAI-compatible HTTP clients (`hull_core.providers.openai_spec`).
76
76
 
77
- Per-task model chains, CSV `provider/model,provider/model`, order = fallback. Provider is inferred from the model prefix.
78
-
79
- - `EMBEDDING_MODELS` -- chain embedding. Empty = local ONNX from the fastretrieval built-in registry.
80
- - `SUMMARY_MODELS` -- chain summarizer (graph `summarize` action). Empty = summaries disabled.
77
+ - `EMBEDDING_MODELS` -- CSV `provider/model,...` selection cho cloud embedding chain; entry dau tien active, cac entry sau la config luu (khong fallback runtime). Empty = local ONNX tu fastretrieval built-in registry.
78
+ - Summary model lay tu `model` field cua `[models.chat]` cell; cell chua cau hinh = summaries disabled. Summary-chain env cu khong con duoc doc.
81
79
  - **Local (default)**: fastretrieval ONNX registry -- zero-config, ~570MB download on first use, 768-dim MRL truncation
82
- - API key theo convention `<PROVIDER>_API_KEY`. 7 provider servers goi y:
83
-
84
- | model prefix | key env var | get it at |
85
- |---|---|---|
86
- | `gemini/` | `GEMINI_API_KEY` | aistudio.google.com/apikey |
87
- | `openai/` (or bare) | `OPENAI_API_KEY` | platform.openai.com |
88
- | `jina_ai/` | `JINA_AI_API_KEY` | jina.ai/api-key |
89
- | `cohere/` | `COHERE_API_KEY` | dashboard.cohere.com |
90
- | `xai/` | `XAI_API_KEY` | console.x.ai |
91
- | `anthropic/` | `ANTHROPIC_API_KEY` | console.anthropic.com |
92
- | `vertex_express/` | `GOOGLE_VERTEX_EXPRESS_API_KEY` | cloud.google.com/vertex-ai/generative-ai/docs/start/express-mode/overview |
93
-
94
- Summarizer providers must expose a chat-completion API (Jina/Cohere do not).
95
- - Custom endpoint (SSRF-guarded): `EMBEDDING_API_BASE` (embedding), `LLM_API_BASE` (summarizer)
80
+ - Transport + credential lay tu per-task `[models.<task>]` cell trong instance config (`$CRG_CONFIG_DIR` hoac `~/.crg/config.toml`): `base_url` + `api_key` + `model`, plain OpenAI-spec HTTP qua hull-core, OpenRouter pre-wired default (`hull config init`). Env override: `HULL_EMBED_API_KEY` / `HULL_CHAT_API_KEY`.
81
+ - Model-name prefix (`cohere/…`, `jina_ai/…`, `gemini/…`, `openrouter/…`) chi chon wire detail (vd Cohere `input_type`); khong chon key hay transport. Khong con per-vendor API-key env — ambient `*_API_KEY` vars are not read.
96
82
  - `DISABLE_LOCAL_EMBED` -- skip local ONNX download; `resolve_backend` returns `unavailable` (not local) when no cloud chain is configured
97
83
  - Fixed 768-dim storage keeps the table schema valid across providers. Switching embedding MODEL changes the vector space; embeddings are tagged per provider and the cosine search restricts to the active provider, so a provider switch re-embeds rather than mixing incomparable vectors.
98
- - Deprecated (honored one release voi warning): singular `EMBEDDING_MODEL`/`SUMMARY_MODEL` + `EMBEDDING_BACKEND` (backend gio suy ra tu chain rong hay khong). Router auto-detect cu "Jina > Gemini > OpenAI > Cohere" da bo.
84
+ - Deprecated (honored one release voi warning): singular `EMBEDDING_MODEL` + `EMBEDDING_BACKEND` (backend gio suy ra tu chain rong hay khong). Router auto-detect cu "Jina > Gemini > OpenAI > Cohere" da bo. Singular-summary-model va base-URL env vars la dead strings — khong con duoc doc.
99
85
 
100
86
  ### BYO local embedding
101
87
 
@@ -111,20 +97,18 @@ chối, không tự rơi về model mặc định.
111
97
 
112
98
  ### Manual config example
113
99
 
114
- ```json
115
- {
116
- "mcpServers": {
117
- "crg": {
118
- "command": "uvx", "args": ["better-code-review-graph"],
119
- "env": {
120
- "EMBEDDING_MODELS": "jina_ai/jina-embeddings-v5-text-small,gemini/gemini-embedding-001",
121
- "SUMMARY_MODELS": "gemini/gemini-2.5-flash",
122
- "JINA_AI_API_KEY": "jina_xxx",
123
- "GEMINI_API_KEY": "AIza_xxx"
124
- }
125
- }
126
- }
127
- }
100
+ `~/.crg/config.toml`:
101
+
102
+ ```toml
103
+ [models.embed]
104
+ base_url = "https://openrouter.ai/api/v1"
105
+ api_key = "sk-or-..."
106
+ model = "jina-ai/jina-embeddings-v5-text-small"
107
+
108
+ [models.chat]
109
+ base_url = "https://openrouter.ai/api/v1"
110
+ api_key = "sk-or-..."
111
+ model = "minimax/minimax-m3:free"
128
112
  ```
129
113
 
130
114
  ## Pytest
@@ -11,13 +11,16 @@
11
11
  FROM ghcr.io/astral-sh/uv:python3.13-bookworm-slim@sha256:531f855bda2c73cd6ef67d56b733b357cea384185b3022bd09f05e002cd144ca AS builder
12
12
  WORKDIR /app
13
13
  COPY pyproject.toml uv.lock README.md ./
14
- RUN sed -i '/^\[tool\.uv\.sources\]/,/^$/d' pyproject.toml && cp uv.lock /tmp/uv.lock.docker
14
+ # Keep [tool.uv.sources]: uv sync --frozen resolves hull-core from the git
15
+ # source recorded in uv.lock through it. (A previous sed stripped this
16
+ # section — and, on CRLF checkouts, never terminated and ate
17
+ # [project.scripts] + [tool.hatch...force-include] too, producing a wheel
18
+ # without console scripts or bundled migrations/rules.)
15
19
  RUN --mount=type=cache,target=/root/.cache/uv \
16
20
  uv sync --frozen --no-install-project --no-dev
17
21
  COPY src/ src/
18
22
  COPY migrations/ migrations/
19
23
  COPY rules/ rules/
20
- RUN sed -i '/^\[tool\.uv\.sources\]/,/^$/d' pyproject.toml && cp /tmp/uv.lock.docker uv.lock
21
24
  RUN --mount=type=cache,target=/root/.cache/uv \
22
25
  uv sync --frozen --no-dev --no-editable
23
26
 
@@ -49,7 +52,12 @@ ENTRYPOINT ["python", "-m", "crg"]
49
52
  # Stage 3b: http target (multi-user remote daemon)
50
53
  # ========================
51
54
  FROM runtime AS http
55
+ # MCP_HOST=0.0.0.0 so the container is reachable; the shared run_http()
56
+ # loopback guard then requires the mounted instance config to set
57
+ # [server] auth = "token"|"multi" (no-auth + off-loopback bind refuses
58
+ # to start with rc 2 instead of serving unauthenticated).
52
59
  ENV MCP_TRANSPORT=http \
53
- MCP_PORT=8080
60
+ MCP_PORT=8080 \
61
+ MCP_HOST=0.0.0.0
54
62
  EXPOSE 8080
55
- ENTRYPOINT ["crg", "serve"]
63
+ ENTRYPOINT ["crg", "server", "start"]
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.5
2
2
  Name: better-code-review-graph
3
- Version: 3.29.8
3
+ Version: 3.30.0
4
4
  Summary: Token-efficient code review knowledge graph: semantic search and call-graph resolution.
5
5
  Author: n24q02m
6
6
  License-Expression: Apache-2.0
@@ -33,14 +33,14 @@ Description-Content-Type: text/markdown
33
33
 
34
34
  # Better Code Review Graph
35
35
 
36
- > **Renamed (2026-09-13):** repo is now `crg` — CLI-first (`crg` command). PyPI package stays `crg`; MCP server is a secondary surface.
36
+ > **Renamed (2026-09-13):** repo is now `crg` — CLI-first (`crg` command). PyPI package stays `better-code-review-graph`; MCP server is a secondary surface.
37
37
 
38
38
  mcp-name: io.github.n24q02m/crg
39
39
 
40
40
  **Knowledge graph for token-efficient code reviews -- semantic search and call-graph resolution across your codebase.**
41
41
 
42
42
  <!-- Badge Row 1: Status -->
43
- [![Mode](https://img.shields.io/badge/mode-daemon_%C2%B7_http_remote_relay-5C6BC0)](https://mcp.n24q02m.com/get-started/modes-overview/)
43
+ [![Mode](https://img.shields.io/badge/mode-stdio_%C2%B7_http_self_host-5C6BC0)](https://mcp.n24q02m.com/get-started/modes-overview/)
44
44
  [![CI](https://github.com/n24q02m/crg/actions/workflows/ci.yml/badge.svg)](https://github.com/n24q02m/crg/actions/workflows/ci.yml)
45
45
  [![codecov](https://codecov.io/gh/n24q02m/crg/graph/badge.svg)](https://codecov.io/gh/n24q02m/crg)
46
46
  [![PyPI](https://img.shields.io/pypi/v/better-code-review-graph?logo=pypi&logoColor=white)](https://pypi.org/project/better-code-review-graph/)
@@ -96,6 +96,7 @@ v2.0 adds temporal columns (`valid_from_sha` / `valid_to_sha` on every node + ed
96
96
 
97
97
  - [v2.0 migration (BREAKING)](#v20-migration-breaking)
98
98
  - [Install](#install)
99
+ - [Usage](#usage)
99
100
  - [Smithery](#smithery)
100
101
  - [Configuration](#configuration)
101
102
  - [Tools](#tools)
@@ -168,6 +169,69 @@ Install matrix (stdio unless noted; the CLI-first usage above stays the primary
168
169
  Full CLI usage is in [CLI](#cli). Optional per-client MCP setup is at
169
170
  **[mcp.n24q02m.com/servers/better-code-review-graph/setup/](https://mcp.n24q02m.com/servers/better-code-review-graph/setup/)**.
170
171
 
172
+ ## Usage
173
+
174
+ Two ways to run the server, plus the surfaces to consume it.
175
+
176
+ ### Dev: uv, no auth (loopback only)
177
+
178
+ ```bash
179
+ # instance config lives at $CRG_CONFIG_DIR or ~/.crg (same schema as hull's
180
+ # config.toml); a default no-auth template is fine for local dev
181
+ uv run better-code-review-graph server start
182
+ # -> http://127.0.0.1:8000/mcp (no-auth is refused on any off-loopback bind)
183
+ ```
184
+
185
+ ### Always-on: docker (token auth)
186
+
187
+ ```bash
188
+ cp docker-config/config.example.toml docker-config/config.toml
189
+ # mint a token + hash (hull CLI ships with crg's dependency tree):
190
+ # TOKEN=$(python -c "import secrets; print(secrets.token_urlsafe(32))")
191
+ # hull token hash "$TOKEN" # paste output into token_hash in config.toml
192
+ docker compose up -d
193
+ ```
194
+
195
+ The compose file builds the Dockerfile's `http` target (image `crg:http`),
196
+ publishes the container on host loopback only
197
+ (`127.0.0.1:${CRG_PORT:-8772}:8080`), and mounts
198
+ `./docker-config/config.toml` read-only at the container's config dir
199
+ (`CRG_CONFIG_DIR=/data/config`); graph state persists in the `crg-data`
200
+ named volume. All state stays on your machine.
201
+
202
+ ### Consuming: CLI or MCP
203
+
204
+ CLI (local graphs, no server needed):
205
+
206
+ ```bash
207
+ crg graph build --full-rebuild --repo-root /path/to/repo
208
+ crg query search --search-query "authentication" --repo-root /path/to/repo
209
+ ```
210
+
211
+ MCP over HTTP (remote/always-on): point any MCP client at
212
+ `http://127.0.0.1:8772/mcp` with `Authorization: Bearer <token>`:
213
+
214
+ ```json
215
+ {
216
+ "mcpServers": {
217
+ "better-code-review-graph": {
218
+ "url": "http://127.0.0.1:8772/mcp",
219
+ "headers": { "Authorization": "Bearer <token>" }
220
+ }
221
+ }
222
+ }
223
+ ```
224
+
225
+ ### Per-task model configuration
226
+
227
+ Each task (embed / rerank / chat / jev_score) resolves its own OpenAI-spec
228
+ provider cell from `[models.<task>]` in the instance config: an independent
229
+ `base_url` + `api_key` + `model`. Cloud or local is purely a config choice —
230
+ point `base_url` at OpenRouter, a vendor, or your own local gateway
231
+ (e.g. `http://host.docker.internal:11434/v1`). API keys are host-only
232
+ material: keep them in `config.toml` (gitignored under `docker-config/`) or
233
+ supply via `HULL_<TASK>_API_KEY` env; they are never end-user supplied.
234
+
171
235
  ## Local-first boundary
172
236
 
173
237
  CRG is local-first for coding workflows:
@@ -212,14 +276,16 @@ for cloud embeddings, LLM summaries, or an explicit BYO local artifact.
212
276
  ### Model selection
213
277
 
214
278
  Embeddings select the first `provider/model` entry in `EMBEDDING_MODELS`; later
215
- entries are retained as configuration but are not runtime fallbacks. Summaries
216
- select the first `SUMMARY_MODELS` entry too, without runtime fallback. Providers
217
- are inferred from model prefixes and use the matching `<PROVIDER>_API_KEY`.
279
+ entries are retained as configuration but are not runtime fallbacks. The summary
280
+ model comes from the `model` field of the `[models.chat]` cell (an unconfigured
281
+ cell disables summaries). Transport always comes from the per-task
282
+ `[models.<task>]` cell (OpenRouter pre-wired default) — model-name prefixes
283
+ only select wire details, never keys.
218
284
 
219
285
  | Variable | Purpose | Empty (default) |
220
286
  |---|---|---|
221
287
  | `EMBEDDING_MODELS` | Cloud embedding selection; the first entry is active | Local fastretrieval registry |
222
- | `SUMMARY_MODELS` | Completion model selection for `graph(action="summarize")` | Summaries disabled |
288
+ | `[models.chat]` cell `model` | Completion model selection for `graph(action="summarize")` | Summaries disabled |
223
289
 
224
290
  Cohere `embed-v4.0` requests and stores **1024 dimensions**; other backends retain
225
291
  768-dimensional storage. CRG never slices, pads, or silently accepts a different
@@ -230,35 +296,33 @@ nodes are retained and re-embedding replaces only stale vectors.
230
296
 
231
297
  ### Provider API keys
232
298
 
233
- Cloud models need the provider key for the selected model prefix. Keys alone
234
- never select models: an empty embedding chain stays local, and an empty summary
299
+ The key lives in the per-task cell's `api_key` (host-only; see the table below).
300
+ Keys alone never select models: an empty embedding chain stays local, and an empty summary
235
301
  chain stays disabled. A configured cloud error does not fall back to local or
236
302
  another provider. Summarizers require a chat-completion model.
237
303
 
238
- | Model prefix | API key env var | Get a key |
304
+ | Task | Config cell | Notes |
239
305
  |---|---|---|
240
- | `jina_ai/` | `JINA_AI_API_KEY` | <https://jina.ai/api-key> |
241
- | `gemini/` | `GEMINI_API_KEY` (or `GOOGLE_API_KEY`) | <https://aistudio.google.com/apikey> |
242
- | `openai/` (or bare `text-embedding-*`) | `OPENAI_API_KEY` | <https://platform.openai.com/api-keys> |
243
- | `cohere/` | `COHERE_API_KEY` | <https://dashboard.cohere.com/api-keys> |
244
- | `openrouter/` | `OPENROUTER_API_KEY` | <https://openrouter.ai/settings/keys> |
245
- | `vertex_express/` | `GOOGLE_VERTEX_EXPRESS_API_KEY` | <https://cloud.google.com/vertex-ai/generative-ai/docs/start/express-mode/overview> |
306
+ | Embedding | `[models.embed]` in the instance config (`base_url` + `api_key` + `model`), or `HULL_EMBED_API_KEY` env | Plain OpenAI-spec HTTP through hull-core; OpenRouter pre-wired default. `EMBEDDING_MODELS` picks the model name(s) — keys never select a model. Model prefixes (`cohere/…`, `jina_ai/…`, `gemini/…`) select wire details only (e.g. Cohere `input_type`); there are no per-vendor API keys. Cohere embed-4.0 direct is the documented opt-in exception (spec 2026-09-26 §9). |
307
+ | Summarizer | `[models.chat]` cell (or `HULL_CHAT_API_KEY`) | Requires a chat-completion model; OpenRouter default. |
308
+
309
+ Vendor-specific key env vars (`JINA_AI_API_KEY`, `GEMINI_API_KEY`,
310
+ `OPENAI_API_KEY`, `GOOGLE_VERTEX_EXPRESS_API_KEY`, …) are **not read** by
311
+ crg; the removed per-vendor dispatch lived in the pre-de-host stack.
246
312
 
247
313
  ### Advanced
248
314
 
249
315
  | Variable | Purpose |
250
316
  |---|---|
251
- | `EMBEDDING_API_BASE` | Provider-compatible endpoint for cloud embedding, including CF AI Gateway (SSRF-guarded) |
252
- | `LLM_API_BASE` | Provider-compatible base URL for the summarizer, including CF AI Gateway (SSRF-guarded) |
317
+ | `EMBEDDING_MODELS` | Comma-separated embedding model chain; the first entry is the cloud model (host-owned; keys never select a model). Legacy `EMBEDDING_MODEL` is honored with a deprecation warning until the next release. |
253
318
  | `DISABLE_LOCAL_EMBED` | Skip the local ONNX download; embedding is unavailable unless a cloud chain is configured |
254
- | `LOCAL_EMBEDDING_MODEL` | Built-in fastretrieval model ID, or a local directory containing `fastretrieval-manifest.json` | Built-in default |
255
319
  | `LOCAL_RERANK_MODEL` | Fastretrieval `TextCrossEncoder` model ID for bounded semantic reranking | Blank (disabled) |
256
320
  | `LOCAL_EMBEDDING_DIM` | Required dimension for an external model ID without a manifest | `0` |
257
321
  | `LOCAL_EMBEDDING_MODEL_FILE` | ONNX file path inside a manifest-backed artifact directory | `onnx/model.onnx` |
258
322
  | `LOCAL_EMBEDDING_POOLING` | Explicit pooling for an external ID without a manifest: `CLS`, `MEAN`, `LAST_TOKEN`, or `DISABLED` | `MEAN` |
259
323
  | `LOCAL_EMBEDDING_NORMALIZE` | Explicit L2 normalization for an external ID without a manifest | `true` |
260
324
  | `CRG_DATA_DIR` | Override the per-user data directory (default `~/.crg`) used for per-user graphs and credentials in HTTP multi-user mode |
261
- | `EMBEDDING_BACKEND` / `EMBEDDING_MODEL` / `SUMMARY_MODEL` | **Deprecated** singular vars, honored one release with a warning -- migrate to the `*_MODELS` chains |
325
+ | `EMBEDDING_BACKEND` / `EMBEDDING_MODEL` | **Deprecated** singular vars, honored one release with a warning -- migrate to `EMBEDDING_MODELS`. The pre-de-host summary-model and base-URL env vars are no longer read; model + transport live in the `[models.*]` cells |
262
326
 
263
327
  When `LOCAL_RERANK_MODEL` is configured, semantic vector search retrieves a
264
328
  bounded candidate pool of `min(max(limit * 4, limit), 100)` rows, applies the
@@ -272,6 +336,24 @@ not invoke the reranker.
272
336
 
273
337
  ### Example -- cloud embeddings + summaries
274
338
 
339
+ `~/.crg/config.toml` (write by hand, or start from `hull config init` which
340
+ pre-wires OpenRouter defaults):
341
+
342
+ ```toml
343
+ [models.embed]
344
+ base_url = "https://openrouter.ai/api/v1"
345
+ api_key = "sk-or-..."
346
+ model = "cohere/embed-v4.0"
347
+
348
+ [models.chat]
349
+ base_url = "https://openrouter.ai/api/v1"
350
+ api_key = "sk-or-..."
351
+ model = "minimax/minimax-m3:free"
352
+ ```
353
+
354
+ Equivalent env surface for the MCP server entry (keys override the cells;
355
+ model names still come from `EMBEDDING_MODELS` / the `model` field):
356
+
275
357
  ```json
276
358
  {
277
359
  "mcpServers": {
@@ -281,11 +363,8 @@ not invoke the reranker.
281
363
  "env": {
282
364
  "MCP_TRANSPORT": "stdio",
283
365
  "EMBEDDING_MODELS": "cohere/embed-v4.0",
284
- "SUMMARY_MODELS": "openrouter/minimax/minimax-m3:free",
285
- "EMBEDDING_API_BASE": "https://gateway.ai.cloudflare.com/v1/<account>/<gateway>/cohere/v2/embed",
286
- "LLM_API_BASE": "https://gateway.ai.cloudflare.com/v1/<account>/<gateway>/openrouter/v1",
287
- "COHERE_API_KEY": "<cohere-key>",
288
- "OPENROUTER_API_KEY": "<openrouter-key>"
366
+ "HULL_EMBED_API_KEY": "sk-or-...",
367
+ "HULL_CHAT_API_KEY": "sk-or-..."
289
368
  }
290
369
  }
291
370
  }
@@ -293,9 +372,9 @@ not invoke the reranker.
293
372
  ```
294
373
 
295
374
  Cohere embedding is paid. Authorize a bounded budget before a live index/query;
296
- the Minimax-free completion choice does not make embeddings free. This example
297
- does not add a process-wide model override: missing subject credentials fail
298
- closed rather than inheriting the server environment.
375
+ the Minimax-free completion choice does not make embeddings free. Model-name
376
+ prefixes (`cohere/…`, `openrouter/…`) select wire details only — transport and
377
+ credentials come from the `[models.*]` cells, which the host owns.
299
378
 
300
379
  CRG currently has **no cloud rerank call**: `LOCAL_RERANK_MODEL` is its only
301
380
  reranking path. Setting `RERANK_MODELS` or `RERANK_API_BASE` does not enable one.
@@ -315,7 +394,7 @@ Actions: `build` | `update` | `stats` | `embed` | `export` | `summarize`
315
394
  | `stats` | Graph size, languages, node/edge breakdown, embedding count. |
316
395
  | `embed` | Compute vector embeddings for semantic search. Dual-mode: local ONNX or cloud chain. |
317
396
  | `export` | Export the graph as `graphml` / `json-ld` / `dot` / `cypher`. Inline or to `output_path`. |
318
- | `summarize` | LLM-generated one-paragraph docstrings for `Function` nodes (via the first explicit `SUMMARY_MODELS` entry; no-op when no model is selected). Calls bounded by `max_nodes`. |
397
+ | `summarize` | LLM-generated one-paragraph docstrings for `Function` nodes (via the `[models.chat]` cell; no-op when no model is configured). Calls bounded by `max_nodes`. |
319
398
 
320
399
  ### `query` -- Graph queries
321
400
 
@@ -457,7 +536,7 @@ Sources: [Greptile](https://www.greptile.com/docs/introduction) · [Greptile pri
457
536
  - **Explicit selection** -- Cloud embedding errors are reported; the runtime does not silently switch models or fall back to local ONNX.
458
537
  - **Error handling** -- Tools return error strings with fix suggestions, never crash.
459
538
  - **Read-only mount** -- Docker mode mounts the repo as `:ro` (read-only).
460
- - **SSRF-guarded endpoints** -- Custom `EMBEDDING_API_BASE` / `LLM_API_BASE` URLs are validated before any outbound call.
539
+ - **SSRF-guarded endpoints** -- Custom `[models.*]` cell `base_url` values are validated before any outbound call.
461
540
 
462
541
  To report a vulnerability, see [SECURITY.md](SECURITY.md).
463
542