onto-mcp 0.3.2 → 0.4.1
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.
- package/.onto/processes/reconstruct/actionable-ontology-seed-recomposition-design.md +447 -0
- package/.onto/processes/reconstruct/foundry-style-ontology-seed-contract.md +934 -0
- package/.onto/processes/reconstruct/reconstruct-boundary-contract.md +303 -725
- package/.onto/processes/reconstruct/reconstruct-contract-registry.yaml +1645 -0
- package/.onto/processes/reconstruct/reconstruct-execution-ux-contract.md +26 -22
- package/.onto/processes/reconstruct/source-profile-contract.md +49 -23
- package/.onto/processes/reconstruct/source-profiles/code.md +6 -3
- package/.onto/processes/reconstruct/source-profiles/database.md +5 -2
- package/.onto/processes/reconstruct/source-profiles/document.md +5 -2
- package/.onto/processes/reconstruct/source-profiles/spreadsheet.md +5 -4
- package/.onto/processes/review/review-execution-ux-contract.md +40 -0
- package/.onto/processes/shared/pipeline-execution-ledger-contract.md +26 -10
- package/.onto/processes/shared/target-material-kind-contract.md +29 -16
- package/AGENTS.md +6 -4
- package/README.md +149 -76
- package/dist/cli.js +8 -8
- package/dist/core-api/reconstruct-api.js +117 -31
- package/dist/core-api/review-api.js +47 -0
- package/dist/core-runtime/cli/codex-review-unit-executor.js +39 -2
- package/dist/core-runtime/cli/complete-review-session.js +2 -2
- package/dist/core-runtime/cli/mock-review-unit-executor.js +1 -1
- package/dist/core-runtime/cli/review-invoke.js +9 -9
- package/dist/core-runtime/cli/run-review-prompt-execution.js +39 -5
- package/dist/core-runtime/cli/spawn-watcher.js +266 -47
- package/dist/core-runtime/cli/start-review-session.js +3 -3
- package/dist/core-runtime/llm/llm-caller.js +11 -0
- package/dist/core-runtime/llm/llm-tool-loop.js +2 -0
- package/dist/core-runtime/observability/runtime-stream-observation.js +118 -0
- package/dist/core-runtime/onboard/cli-host.js +174 -0
- package/dist/core-runtime/onboard/host-target.js +22 -0
- package/dist/core-runtime/onboard/json-config-host.js +122 -0
- package/dist/core-runtime/onboard/path-scan.js +26 -0
- package/dist/core-runtime/onboard/prompt.js +51 -0
- package/dist/core-runtime/onboard/register.js +214 -0
- package/dist/core-runtime/onboard/types.js +27 -0
- package/dist/core-runtime/reconstruct/actionable-seed-validation.js +1777 -0
- package/dist/core-runtime/reconstruct/artifact-types.js +10 -4
- package/dist/core-runtime/reconstruct/contract-registry.js +623 -0
- package/dist/core-runtime/reconstruct/domain-id.js +10 -0
- package/dist/core-runtime/reconstruct/governing-snapshot.js +716 -0
- package/dist/core-runtime/reconstruct/material-profile-validation.js +191 -0
- package/dist/core-runtime/reconstruct/materialize-preparation.js +49 -11
- package/dist/core-runtime/reconstruct/pipeline-execution-ledger.js +269 -79
- package/dist/core-runtime/reconstruct/post-seed-validation.js +1194 -51
- package/dist/core-runtime/reconstruct/record.js +104 -20
- package/dist/core-runtime/reconstruct/run.js +2107 -413
- package/dist/core-runtime/reconstruct/seed-claim-projections.js +268 -0
- package/dist/core-runtime/reconstruct/source-profiles.js +93 -4
- package/dist/core-runtime/reconstruct/terminal-validation.js +807 -0
- package/dist/core-runtime/review/review-invocation-runner.js +4 -4
- package/dist/mcp/server.js +110 -38
- package/dist/mcp/tool-schemas.js +20 -6
- package/package.json +8 -17
- package/scripts/onto-review-watch.sh +486 -0
- package/scripts/onto-runtime-watch.sh +122 -0
- package/scripts/postinstall-hint.js +22 -0
- package/.onto/processes/reconstruct/top-level-concept-discovery-contract.md +0 -387
- package/dist/core-runtime/cli/bootstrap-review-binding.js +0 -186
- package/dist/core-runtime/cli/codex-nested-dispatch.test.js +0 -390
- package/dist/core-runtime/cli/codex-nested-teamlead-executor.test.js +0 -335
- package/dist/core-runtime/cli/coordinator-helpers.js +0 -583
- package/dist/core-runtime/cli/coordinator-state-machine-deliberation.test.js +0 -167
- package/dist/core-runtime/cli/coordinator-state-machine.js +0 -794
- package/dist/core-runtime/cli/e2e-codex-multi-agent-fixes.test.js +0 -615
- package/dist/core-runtime/cli/e2e-start-review-session.test.js +0 -312
- package/dist/core-runtime/cli/health.js +0 -44
- package/dist/core-runtime/cli/inline-http-review-unit-executor.test.js +0 -567
- package/dist/core-runtime/cli/materialize-review-execution-preparation.js +0 -104
- package/dist/core-runtime/cli/migrate-session-roots.js +0 -118
- package/dist/core-runtime/cli/repo-layout-migration-replace.smoke.test.js +0 -106
- package/dist/core-runtime/cli/review-invoke-auto-resolution.test.js +0 -268
- package/dist/core-runtime/cli/review-invoke-coordinator-topology.test.js +0 -136
- package/dist/core-runtime/cli/review-invoke-resolver-caching.test.js +0 -201
- package/dist/core-runtime/cli/review-invoke-topology-dispatch.test.js +0 -192
- package/dist/core-runtime/cli/session-root-guard.js +0 -168
- package/dist/core-runtime/cli/spawn-watcher.test.js +0 -457
- package/dist/core-runtime/cli/strip-wrapping-code-fence.test.js +0 -79
- package/dist/core-runtime/cli/teamcreate-lens-deliberation-executor.js +0 -412
- package/dist/core-runtime/cli/teamcreate-lens-deliberation-executor.test.js +0 -351
- package/dist/core-runtime/cli/topology-executor-mapping.js +0 -139
- package/dist/core-runtime/cli/topology-executor-mapping.test.js +0 -173
- package/dist/core-runtime/cli/write-review-interpretation.js +0 -81
- package/dist/core-runtime/config/onto-config-cli.js +0 -278
- package/dist/core-runtime/config/onto-config-key-path.js +0 -288
- package/dist/core-runtime/config/onto-config-key-path.test.js +0 -195
- package/dist/core-runtime/config/onto-config-preview.js +0 -108
- package/dist/core-runtime/config/onto-config-preview.test.js +0 -132
- package/dist/core-runtime/discovery/config-chain.js +0 -118
- package/dist/core-runtime/discovery/config-chain.test.js +0 -103
- package/dist/core-runtime/discovery/config-profile.js +0 -199
- package/dist/core-runtime/discovery/config-profile.test.js +0 -233
- package/dist/core-runtime/discovery/host-detection.test.js +0 -186
- package/dist/core-runtime/discovery/installation-paths.test.js +0 -65
- package/dist/core-runtime/discovery/lens-registry.test.js +0 -81
- package/dist/core-runtime/discovery/path-normalization.test.js +0 -22
- package/dist/core-runtime/discovery/plugin-path.js +0 -72
- package/dist/core-runtime/discovery/plugin-path.test.js +0 -95
- package/dist/core-runtime/evolve/adapters/code-product/compile/compile-defense.js +0 -344
- package/dist/core-runtime/evolve/adapters/code-product/compile/compile-defense.test.js +0 -915
- package/dist/core-runtime/evolve/adapters/code-product/compile/compile.js +0 -564
- package/dist/core-runtime/evolve/adapters/code-product/compile/compile.test.js +0 -708
- package/dist/core-runtime/evolve/adapters/code-product/parsers/brief-parser.js +0 -165
- package/dist/core-runtime/evolve/adapters/code-product/parsers/brief-parser.test.js +0 -227
- package/dist/core-runtime/evolve/adapters/code-product/validators/validate.js +0 -59
- package/dist/core-runtime/evolve/adapters/code-product/validators/validate.test.js +0 -205
- package/dist/core-runtime/evolve/adapters/methodology/adapter.js +0 -16
- package/dist/core-runtime/evolve/adapters/methodology/adapter.test.js +0 -9
- package/dist/core-runtime/evolve/adapters/methodology/perspectives/authority-consistency.js +0 -298
- package/dist/core-runtime/evolve/adapters/methodology/perspectives/authority-consistency.test.js +0 -70
- package/dist/core-runtime/evolve/adapters/methodology/scope-types/process.js +0 -46
- package/dist/core-runtime/evolve/adapters/methodology/scope-types/process.test.js +0 -73
- package/dist/core-runtime/evolve/adapters/registry.js +0 -47
- package/dist/core-runtime/evolve/adapters/registry.test.js +0 -67
- package/dist/core-runtime/evolve/cli.js +0 -256
- package/dist/core-runtime/evolve/commands/align.js +0 -194
- package/dist/core-runtime/evolve/commands/align.test.js +0 -82
- package/dist/core-runtime/evolve/commands/apply.js +0 -161
- package/dist/core-runtime/evolve/commands/apply.test.js +0 -138
- package/dist/core-runtime/evolve/commands/close.js +0 -39
- package/dist/core-runtime/evolve/commands/close.test.js +0 -99
- package/dist/core-runtime/evolve/commands/defer.js +0 -40
- package/dist/core-runtime/evolve/commands/defer.test.js +0 -134
- package/dist/core-runtime/evolve/commands/draft.js +0 -323
- package/dist/core-runtime/evolve/commands/draft.test.js +0 -178
- package/dist/core-runtime/evolve/commands/e2e-evolve-full-cycle.test.js +0 -208
- package/dist/core-runtime/evolve/commands/error-messages.js +0 -125
- package/dist/core-runtime/evolve/commands/error-messages.test.js +0 -167
- package/dist/core-runtime/evolve/commands/propose-align.js +0 -222
- package/dist/core-runtime/evolve/commands/propose-align.test.js +0 -136
- package/dist/core-runtime/evolve/commands/reconstruct.js +0 -330
- package/dist/core-runtime/evolve/commands/reconstruct.test.js +0 -278
- package/dist/core-runtime/evolve/commands/shared.js +0 -22
- package/dist/core-runtime/evolve/commands/stale-check.js +0 -103
- package/dist/core-runtime/evolve/commands/stale-check.test.js +0 -84
- package/dist/core-runtime/evolve/commands/start.js +0 -887
- package/dist/core-runtime/evolve/commands/start.test.js +0 -396
- package/dist/core-runtime/evolve/config/project-config.js +0 -99
- package/dist/core-runtime/evolve/config/project-config.test.js +0 -170
- package/dist/core-runtime/evolve/renderers/align-packet.js +0 -280
- package/dist/core-runtime/evolve/renderers/align-packet.test.js +0 -332
- package/dist/core-runtime/evolve/renderers/draft-packet.js +0 -303
- package/dist/core-runtime/evolve/renderers/draft-packet.test.js +0 -377
- package/dist/core-runtime/evolve/renderers/format.js +0 -5
- package/dist/core-runtime/evolve/renderers/scope-md.js +0 -237
- package/dist/core-runtime/evolve/renderers/scope-md.test.js +0 -306
- package/dist/core-runtime/govern/cli.js +0 -369
- package/dist/core-runtime/govern/cli.test.js +0 -314
- package/dist/core-runtime/govern/drift-engine.js +0 -103
- package/dist/core-runtime/govern/drift-engine.test.js +0 -319
- package/dist/core-runtime/govern/promote-principle.js +0 -206
- package/dist/core-runtime/govern/promote-principle.test.js +0 -368
- package/dist/core-runtime/govern/queue.js +0 -81
- package/dist/core-runtime/govern/types.js +0 -16
- package/dist/core-runtime/install/cli.js +0 -530
- package/dist/core-runtime/install/detect.js +0 -128
- package/dist/core-runtime/install/detect.test.js +0 -155
- package/dist/core-runtime/install/gitignore-update.js +0 -74
- package/dist/core-runtime/install/gitignore-update.test.js +0 -64
- package/dist/core-runtime/install/install-integration.test.js +0 -373
- package/dist/core-runtime/install/prompts.js +0 -389
- package/dist/core-runtime/install/prompts.test.js +0 -293
- package/dist/core-runtime/install/types.js +0 -26
- package/dist/core-runtime/install/validation.js +0 -295
- package/dist/core-runtime/install/validation.test.js +0 -313
- package/dist/core-runtime/install/writer.js +0 -254
- package/dist/core-runtime/install/writer.test.js +0 -218
- package/dist/core-runtime/learning/extractor.js +0 -461
- package/dist/core-runtime/learning/feedback.js +0 -179
- package/dist/core-runtime/learning/health-report.js +0 -165
- package/dist/core-runtime/learning/health-report.test.js +0 -169
- package/dist/core-runtime/learning/loader.js +0 -388
- package/dist/core-runtime/learning/loader.test.js +0 -102
- package/dist/core-runtime/learning/promote/apply-state.js +0 -240
- package/dist/core-runtime/learning/promote/audit-obligation.js +0 -195
- package/dist/core-runtime/learning/promote/collector.js +0 -432
- package/dist/core-runtime/learning/promote/degraded-state.js +0 -125
- package/dist/core-runtime/learning/promote/domain-doc-proposer.js +0 -166
- package/dist/core-runtime/learning/promote/e2e-promote.test.js +0 -6385
- package/dist/core-runtime/learning/promote/health-snapshot.js +0 -150
- package/dist/core-runtime/learning/promote/insight-reclassifier.js +0 -544
- package/dist/core-runtime/learning/promote/judgment-auditor.js +0 -517
- package/dist/core-runtime/learning/promote/panel-reviewer.js +0 -1158
- package/dist/core-runtime/learning/promote/promote-executor.js +0 -1675
- package/dist/core-runtime/learning/promote/promoter.js +0 -307
- package/dist/core-runtime/learning/promote/retirement.js +0 -122
- package/dist/core-runtime/learning/promote/types.js +0 -23
- package/dist/core-runtime/learning/prompt-sections.js +0 -51
- package/dist/core-runtime/learning/shared/artifact-registry-init.js +0 -45
- package/dist/core-runtime/learning/shared/artifact-registry.js +0 -254
- package/dist/core-runtime/learning/shared/audit-obligation-kernel.js +0 -73
- package/dist/core-runtime/learning/shared/audit-state.js +0 -99
- package/dist/core-runtime/learning/shared/duplicate-check.js +0 -28
- package/dist/core-runtime/learning/shared/llm-caller.js +0 -831
- package/dist/core-runtime/learning/shared/llm-caller.test.js +0 -601
- package/dist/core-runtime/learning/shared/llm-tool-loop.js +0 -393
- package/dist/core-runtime/learning/shared/mode.js +0 -25
- package/dist/core-runtime/learning/shared/paths.js +0 -84
- package/dist/core-runtime/learning/shared/paths.test.js +0 -79
- package/dist/core-runtime/learning/shared/patterns.js +0 -37
- package/dist/core-runtime/learning/shared/recoverability.js +0 -355
- package/dist/core-runtime/learning/shared/recovery-context.js +0 -374
- package/dist/core-runtime/learning/shared/scope.js +0 -1
- package/dist/core-runtime/learning/shared/semantic-classifier.js +0 -94
- package/dist/core-runtime/learning/shared/specs/apply-execution-state-spec.js +0 -42
- package/dist/core-runtime/learning/shared/specs/audit-state-spec.js +0 -37
- package/dist/core-runtime/learning/shared/specs/backup-metadata-spec.js +0 -39
- package/dist/core-runtime/learning/shared/specs/emergency-log-spec.js +0 -41
- package/dist/core-runtime/learning/shared/specs/layout-version-spec.js +0 -38
- package/dist/core-runtime/learning/shared/specs/promote-decisions-spec.js +0 -43
- package/dist/core-runtime/learning/shared/specs/promote-report-spec.js +0 -113
- package/dist/core-runtime/learning/shared/specs/prune-log-spec.js +0 -36
- package/dist/core-runtime/learning/shared/specs/recovery-resolution-spec.js +0 -48
- package/dist/core-runtime/learning/shared/specs/restore-manifest-spec.js +0 -43
- package/dist/core-runtime/learning/shared/specs/spec-helpers.js +0 -64
- package/dist/core-runtime/learning/usage-tracker.js +0 -190
- package/dist/core-runtime/learning/usage-tracker.test.js +0 -176
- package/dist/core-runtime/onboard/detect-review-axes.js +0 -122
- package/dist/core-runtime/onboard/detect-review-axes.test.js +0 -127
- package/dist/core-runtime/onboard/write-review-block.js +0 -188
- package/dist/core-runtime/onboard/write-review-block.test.js +0 -240
- package/dist/core-runtime/readers/brownfield-builder.js +0 -150
- package/dist/core-runtime/readers/brownfield-builder.test.js +0 -136
- package/dist/core-runtime/readers/code-chunk-collector.js +0 -53
- package/dist/core-runtime/readers/code-chunk-collector.test.js +0 -136
- package/dist/core-runtime/readers/file-utils.js +0 -240
- package/dist/core-runtime/readers/file-utils.test.js +0 -146
- package/dist/core-runtime/readers/lexicon-citation-check.js +0 -93
- package/dist/core-runtime/readers/lexicon-citation-check.test.js +0 -77
- package/dist/core-runtime/readers/mcp-figma.js +0 -30
- package/dist/core-runtime/readers/mcp-figma.test.js +0 -82
- package/dist/core-runtime/readers/mcp-generic.js +0 -31
- package/dist/core-runtime/readers/mcp-generic.test.js +0 -76
- package/dist/core-runtime/readers/ontology-index.js +0 -148
- package/dist/core-runtime/readers/ontology-index.test.js +0 -245
- package/dist/core-runtime/readers/ontology-query.js +0 -168
- package/dist/core-runtime/readers/ontology-query.test.js +0 -311
- package/dist/core-runtime/readers/ontology-resolve.js +0 -48
- package/dist/core-runtime/readers/ontology-resolve.test.js +0 -48
- package/dist/core-runtime/readers/patterns/index.js +0 -7
- package/dist/core-runtime/readers/review-log.js +0 -213
- package/dist/core-runtime/readers/review-log.test.js +0 -313
- package/dist/core-runtime/readers/scan-local.js +0 -102
- package/dist/core-runtime/readers/scan-local.test.js +0 -102
- package/dist/core-runtime/readers/scan-tarball.js +0 -121
- package/dist/core-runtime/readers/scan-tarball.test.js +0 -283
- package/dist/core-runtime/readers/scan-vault.js +0 -34
- package/dist/core-runtime/readers/scan-vault.test.js +0 -81
- package/dist/core-runtime/readers/types.js +0 -42
- package/dist/core-runtime/readers/types.test.js +0 -94
- package/dist/core-runtime/readers/viewpoint-collectors.js +0 -229
- package/dist/core-runtime/reconstruct/seed-candidate-validation.js +0 -385
- package/dist/core-runtime/review/citation-audit.test.js +0 -165
- package/dist/core-runtime/review/execution-plan-resolver.js +0 -247
- package/dist/core-runtime/review/execution-plan-resolver.test.js +0 -243
- package/dist/core-runtime/review/execution-topology-resolver-axis-first.test.js +0 -246
- package/dist/core-runtime/review/execution-topology-resolver.js +0 -401
- package/dist/core-runtime/review/execution-topology-resolver.test.js +0 -315
- package/dist/core-runtime/review/inline-context-embedder.test.js +0 -154
- package/dist/core-runtime/review/legacy-mode-policy.js +0 -88
- package/dist/core-runtime/review/materializers-effort-persist.test.js +0 -79
- package/dist/core-runtime/review/ontology-path-classifier.js +0 -179
- package/dist/core-runtime/review/ontology-path-classifier.test.js +0 -216
- package/dist/core-runtime/review/packet-boundary-policy.test.js +0 -107
- package/dist/core-runtime/review/participating-lens-paths.test.js +0 -73
- package/dist/core-runtime/review/review-config-legacy-translate.js +0 -244
- package/dist/core-runtime/review/review-config-legacy-translate.test.js +0 -161
- package/dist/core-runtime/review/review-config-validator.js +0 -289
- package/dist/core-runtime/review/review-config-validator.test.js +0 -236
- package/dist/core-runtime/review/shape-pipeline-audit.test.js +0 -311
- package/dist/core-runtime/review/shape-to-topology-id.js +0 -117
- package/dist/core-runtime/review/shape-to-topology-id.test.js +0 -132
- package/dist/core-runtime/review/topology-shape-derivation.js +0 -155
- package/dist/core-runtime/review/topology-shape-derivation.test.js +0 -195
- package/dist/core-runtime/scope-runtime/constants.js +0 -12
- package/dist/core-runtime/scope-runtime/constraint-pool.js +0 -166
- package/dist/core-runtime/scope-runtime/constraint-pool.test.js +0 -674
- package/dist/core-runtime/scope-runtime/domain-validation-log.js +0 -135
- package/dist/core-runtime/scope-runtime/domain-validation-log.test.js +0 -156
- package/dist/core-runtime/scope-runtime/eval-persistence.js +0 -65
- package/dist/core-runtime/scope-runtime/eval-persistence.test.js +0 -84
- package/dist/core-runtime/scope-runtime/event-pipeline.js +0 -64
- package/dist/core-runtime/scope-runtime/event-pipeline.test.js +0 -450
- package/dist/core-runtime/scope-runtime/event-store.js +0 -39
- package/dist/core-runtime/scope-runtime/event-store.test.js +0 -95
- package/dist/core-runtime/scope-runtime/gate-guard.js +0 -348
- package/dist/core-runtime/scope-runtime/gate-guard.test.js +0 -1047
- package/dist/core-runtime/scope-runtime/hash.js +0 -4
- package/dist/core-runtime/scope-runtime/hash.test.js +0 -33
- package/dist/core-runtime/scope-runtime/id.js +0 -4
- package/dist/core-runtime/scope-runtime/id.test.js +0 -17
- package/dist/core-runtime/scope-runtime/reducer.js +0 -297
- package/dist/core-runtime/scope-runtime/reducer.test.js +0 -759
- package/dist/core-runtime/scope-runtime/scope-manager.js +0 -161
- package/dist/core-runtime/scope-runtime/state-machine.js +0 -309
- package/dist/core-runtime/scope-runtime/state-machine.test.js +0 -704
- package/dist/core-runtime/scope-runtime/types.js +0 -116
- package/dist/core-runtime/scope-runtime/types.test.js +0 -69
- package/dist/core-runtime/translate/render-for-user.js +0 -169
- package/dist/core-runtime/translate/render-for-user.test.js +0 -122
- package/dist/providers/capability-contract.js +0 -1
|
@@ -7,8 +7,8 @@
|
|
|
7
7
|
## 1. Position
|
|
8
8
|
|
|
9
9
|
`reconstruct` can take a long time because the host LLM must read source
|
|
10
|
-
observations, run reconstruct lens judgments, choose the next unjudged
|
|
11
|
-
frontier, propose
|
|
10
|
+
observations, run reconstruct lens judgments, choose the next unjudged
|
|
11
|
+
source frontier, propose seed meaning, assess competency questions, and explain
|
|
12
12
|
unresolved gaps. The user-facing experience must therefore expose new
|
|
13
13
|
information as it is discovered, not only meta status such as "still running".
|
|
14
14
|
|
|
@@ -34,18 +34,18 @@ The brief must include:
|
|
|
34
34
|
test/fixture-only profile if a non-product harness is running, with the
|
|
35
35
|
completion claim allowed for that profile
|
|
36
36
|
- model/provider: provider family or realization label without secrets
|
|
37
|
-
- domain: selected domain, no-domain mode, or pending domain
|
|
37
|
+
- domain: selected domain, no-domain mode, or pending domain competency admission
|
|
38
38
|
- material: `target_material_kind`, target input kind, and unsupported-material
|
|
39
39
|
status if known
|
|
40
40
|
- exploration loop: initial source frontier, round budget if any, and frontier
|
|
41
41
|
validation policy
|
|
42
42
|
- reconstruction direction: what the reconstruct run will try to explain, and
|
|
43
43
|
what it will treat as out of scope
|
|
44
|
-
- ownership boundary: runtime observes and validates; reconstruct
|
|
45
|
-
meaning and source gaps; the user or host confirmation controls
|
|
44
|
+
- ownership boundary: runtime observes and validates; reconstruct lens judgments judge
|
|
45
|
+
meaning and source gaps; the user or host confirmation controls seed
|
|
46
46
|
acceptance
|
|
47
47
|
- skipped or deferred stages that already narrow downstream authority, such as
|
|
48
|
-
live lens judgment, source-frontier exploration, domain
|
|
48
|
+
live lens judgment, source-frontier exploration, domain competency admission,
|
|
49
49
|
user confirmation, or competency-question assessment
|
|
50
50
|
|
|
51
51
|
The opening brief should be declarative. It should not ask the user to approve
|
|
@@ -78,9 +78,12 @@ Progress updates should prefer facts such as:
|
|
|
78
78
|
- per-lens judgment status and newly named semantic gaps
|
|
79
79
|
- source frontier refs requested, accepted, rejected, already observed,
|
|
80
80
|
unsupported, or out of bounds
|
|
81
|
-
- domain
|
|
82
|
-
-
|
|
83
|
-
|
|
81
|
+
- domain competency admission refs and governing snapshot id
|
|
82
|
+
- seed layer count by object, actor, action, workflow, permission, and data
|
|
83
|
+
binding
|
|
84
|
+
- candidate disposition count by disposition
|
|
85
|
+
- ontology-facing handoff readiness, canonical readiness projection, or
|
|
86
|
+
limitation count
|
|
84
87
|
- confirmation state counts
|
|
85
88
|
- competency question count and assessment status counts
|
|
86
89
|
- failure classification counts
|
|
@@ -89,9 +92,10 @@ Progress updates should prefer facts such as:
|
|
|
89
92
|
- current halt reason and reusable artifact refs when halted
|
|
90
93
|
- skipped/deferred stage reason and downstream authority impact
|
|
91
94
|
|
|
92
|
-
Progress updates should avoid pretending that intermediate semantic
|
|
93
|
-
final. Before
|
|
94
|
-
question assessment, quality statements are
|
|
95
|
+
Progress updates should avoid pretending that intermediate semantic content is
|
|
96
|
+
final. Before seed confirmation, seed content is candidate content. Before
|
|
97
|
+
competency-question assessment and handoff validation, quality statements are
|
|
98
|
+
preliminary.
|
|
95
99
|
|
|
96
100
|
## 4. Decision Points
|
|
97
101
|
|
|
@@ -104,9 +108,8 @@ Decision points:
|
|
|
104
108
|
- material kind is unsupported or mixed in a way that changes adapter behavior
|
|
105
109
|
- source frontier expansion would exceed the declared boundary, adapter support,
|
|
106
110
|
or round/cost budget
|
|
107
|
-
- domain
|
|
108
|
-
-
|
|
109
|
-
state
|
|
111
|
+
- domain competency admission changes the interpretation standard
|
|
112
|
+
- seed confirmation requires accepted, rejected, partial, or deferred state
|
|
110
113
|
- unresolved material questions require a final direction: continue, defer, or
|
|
111
114
|
accept with disclosed limits
|
|
112
115
|
|
|
@@ -118,7 +121,7 @@ The host should phrase choices by outcome, not internal jargon.
|
|
|
118
121
|
separate:
|
|
119
122
|
|
|
120
123
|
- confirmed Seed content
|
|
121
|
-
-
|
|
124
|
+
- candidate disposition summary
|
|
122
125
|
- competency question assessment summary
|
|
123
126
|
- material failures and unsafe-to-trust gaps
|
|
124
127
|
- revision proposals and action candidates
|
|
@@ -132,9 +135,10 @@ of `reconstruct-record.yaml`, `reconstruct-run-manifest.yaml`, and the
|
|
|
132
135
|
stage-owned artifacts.
|
|
133
136
|
|
|
134
137
|
Full integral exploration wording is allowed only when the run manifest records
|
|
135
|
-
the full profile and the required exploration, domain
|
|
136
|
-
confirmation, competency-question, assessment, failure, revision, metrics,
|
|
137
|
-
and final-output stages are trusted or explicitly skipped
|
|
138
|
+
the full profile and the required exploration, domain competency admission, Seed,
|
|
139
|
+
confirmation, competency-question, assessment, failure, revision, metrics,
|
|
140
|
+
handoff-validation, and final-output stages are trusted or explicitly skipped
|
|
141
|
+
with trusted reasons.
|
|
138
142
|
|
|
139
143
|
## 6. Halted Or Partial Runs
|
|
140
144
|
|
|
@@ -151,9 +155,9 @@ If a run halts, the output should still be useful:
|
|
|
151
155
|
trusted artifacts
|
|
152
156
|
|
|
153
157
|
A halted run may provide candidate Seed content only if the corresponding
|
|
154
|
-
artifact and validation refs exist. It may not imply
|
|
155
|
-
assessment, failure classification, revision proposal,
|
|
156
|
-
ontology direction unless those artifacts exist.
|
|
158
|
+
artifact and validation refs exist. It may not imply seed confirmation, CQ
|
|
159
|
+
assessment, failure classification, revision proposal, terminal handoff readiness, or
|
|
160
|
+
final ontology direction unless those artifacts exist.
|
|
157
161
|
|
|
158
162
|
## 7. Runtime Payload Expectations
|
|
159
163
|
|
|
@@ -1,6 +1,7 @@
|
|
|
1
1
|
# Reconstruct Source Profile Contract
|
|
2
2
|
|
|
3
|
-
>
|
|
3
|
+
> Contract status: active.
|
|
4
|
+
> Runtime support status authority: `reconstruct-contract-registry.yaml#source_profile_records`.
|
|
4
5
|
> Purpose: define target-material observation contracts for `reconstruct`.
|
|
5
6
|
|
|
6
7
|
## 1. Canonical Seat
|
|
@@ -11,14 +12,14 @@ Source profiles live under:
|
|
|
11
12
|
.onto/processes/reconstruct/source-profiles/
|
|
12
13
|
```
|
|
13
14
|
|
|
14
|
-
|
|
15
|
-
|
|
16
|
-
|
|
15
|
+
`SourceProfileDefinition` is the contract-owned source profile file. A
|
|
16
|
+
`SelectedSourceProfile` is the runtime-owned selection recorded after material
|
|
17
|
+
classification. Neither concept owns semantic interpretation.
|
|
17
18
|
|
|
18
|
-
In the integral exploration design, source profiles
|
|
19
|
-
observation side. Reconstruct
|
|
20
|
-
a validated source frontier, but the profile itself does not
|
|
21
|
-
is semantically important.
|
|
19
|
+
In the integral exploration design, source profiles belong to the runtime
|
|
20
|
+
observation side. Reconstruct lens judgments may ask for additional
|
|
21
|
+
source refs through a validated source frontier, but the profile itself does not
|
|
22
|
+
decide which source is semantically important.
|
|
22
23
|
|
|
23
24
|
Source profiles are keyed by `target_material_kind`, the shared runtime axis
|
|
24
25
|
defined in `.onto/authority/core-lexicon.yaml`. They must not use `source_kind`
|
|
@@ -29,16 +30,25 @@ to mean code, spreadsheet, document, or database because review already uses
|
|
|
29
30
|
The cross-process goal and validation rules for this axis are defined in
|
|
30
31
|
`.onto/processes/shared/target-material-kind-contract.md`.
|
|
31
32
|
|
|
32
|
-
|
|
33
|
-
|
|
34
|
-
|
|
35
|
-
|
|
36
|
-
|
|
37
|
-
|
|
38
|
-
|
|
39
|
-
|
|
40
|
-
|
|
41
|
-
|
|
33
|
+
The current profile record set is owned by
|
|
34
|
+
`.onto/processes/reconstruct/reconstruct-contract-registry.yaml#source_profile_records`.
|
|
35
|
+
That registry is the executable authority for profile ids, definition refs,
|
|
36
|
+
definition hashes, contract status, runtime implementation status, schema
|
|
37
|
+
version, profile version, and migration status. This source-profile contract
|
|
38
|
+
defines what a profile means; it does not independently publish current support
|
|
39
|
+
status.
|
|
40
|
+
|
|
41
|
+
Profile migration continuity is also registry-owned. `source_profile_records`
|
|
42
|
+
must record `supersedes`, `replaced_by`, `split_from`, `split_into`,
|
|
43
|
+
`merged_from`, and `merged_into` so old profile snapshots can be replayed,
|
|
44
|
+
compared, or migrated without guessing how a previous profile id maps to the
|
|
45
|
+
current registry.
|
|
46
|
+
|
|
47
|
+
`contract_status` means whether a profile definition or public material-kind
|
|
48
|
+
contract is authoritative. `runtime_implementation_status` means whether the
|
|
49
|
+
current runtime can execute that profile. A profile can be contract-active while
|
|
50
|
+
its runtime adapter is still planned or unsupported, but that status must be read
|
|
51
|
+
from `source_profile_records`.
|
|
42
52
|
|
|
43
53
|
## 2. Profile Responsibility
|
|
44
54
|
|
|
@@ -52,11 +62,15 @@ A source profile may define:
|
|
|
52
62
|
- scan targets
|
|
53
63
|
- safe frontier-ref shapes for this material kind
|
|
54
64
|
- correct and incorrect observation examples
|
|
55
|
-
-
|
|
65
|
+
- profile-specific unsupported cases that do not claim current runtime status
|
|
66
|
+
- profile-specific `candidate_subkind` and `disposition_detail` qualifiers
|
|
56
67
|
|
|
57
68
|
A source profile must not define rules that convert source structure into
|
|
58
69
|
ontology concepts or rules that choose the next source based on ontology
|
|
59
|
-
meaning.
|
|
70
|
+
meaning. A source profile must not define or override `contract_status`,
|
|
71
|
+
`runtime_implementation_status`, `schema_version`, `profile_version`, or
|
|
72
|
+
`migration_status`; those values and source-profile migration mapping fields
|
|
73
|
+
belong to `source_profile_records`.
|
|
60
74
|
|
|
61
75
|
Examples:
|
|
62
76
|
|
|
@@ -75,7 +89,7 @@ profile may guide observation scope, but the adapter schema is the runtime
|
|
|
75
89
|
contract that fixes returned fields and observation ids.
|
|
76
90
|
|
|
77
91
|
When a source adapter is invoked after the first round, it consumes only
|
|
78
|
-
runtime-validated source frontier refs. It must not accept lens prose or
|
|
92
|
+
runtime-validated source frontier refs. It must not accept lens-judgment prose or
|
|
79
93
|
semantic labels as source locations.
|
|
80
94
|
|
|
81
95
|
The future adapter contract must fail explicitly when:
|
|
@@ -117,13 +131,25 @@ validates it before any additional observation occurs.
|
|
|
117
131
|
Adding a new target material kind requires:
|
|
118
132
|
|
|
119
133
|
1. A new source profile under `.onto/processes/reconstruct/source-profiles/`.
|
|
120
|
-
2.
|
|
134
|
+
2. A `source_profile_records` entry that declares profile id, definition ref or
|
|
135
|
+
explicit null-ref behavior, definition hash, contract status, runtime
|
|
136
|
+
implementation status, schema version, profile version, migration status,
|
|
137
|
+
and source-profile migration mapping fields.
|
|
121
138
|
3. Tests for target material detection, observation shape, source frontier
|
|
122
139
|
validation, unsupported inputs, and directive evidence-ref validation.
|
|
123
|
-
4.
|
|
140
|
+
4. `reconstruct-contract-registry.yaml` updates when artifact authority,
|
|
141
|
+
validation gates, root candidate kinds, root dispositions, material kinds,
|
|
142
|
+
source profile records, source profile definition refs, runtime
|
|
143
|
+
implementation status values, or support-status migration behavior change.
|
|
144
|
+
5. MCP schema updates only after the runtime contract is implemented.
|
|
124
145
|
|
|
125
146
|
The source profile alone does not make a target material kind supported.
|
|
126
147
|
|
|
148
|
+
Profile-specific refinements must use qualifiers. A source profile may introduce
|
|
149
|
+
`candidate_subkind` or `disposition_detail` values for its material kind. It may
|
|
150
|
+
not introduce a new root candidate kind, root disposition, or material kind
|
|
151
|
+
without a contract and registry change.
|
|
152
|
+
|
|
127
153
|
## 6. Mixed Material Rule
|
|
128
154
|
|
|
129
155
|
`mixed` is a public `TargetMaterialKind` value, but it is not a material parser.
|
|
@@ -2,9 +2,12 @@
|
|
|
2
2
|
|
|
3
3
|
> Target material kind: `code`
|
|
4
4
|
|
|
5
|
-
##
|
|
5
|
+
## Registry Record
|
|
6
6
|
|
|
7
|
-
|
|
7
|
+
Current profile id, contract status, runtime implementation status, schema
|
|
8
|
+
version, profile version, migration status, and definition hash are owned by
|
|
9
|
+
`.onto/processes/reconstruct/reconstruct-contract-registry.yaml#source_profile_records`.
|
|
10
|
+
This file defines code observation semantics only.
|
|
8
11
|
|
|
9
12
|
## Target Material Identification Hints
|
|
10
13
|
|
|
@@ -59,7 +62,7 @@ status field definition -- src/payment.ts:14
|
|
|
59
62
|
- What user-visible workflow does this code support?
|
|
60
63
|
- Is this code a service, library, script, UI, integration, or test harness?
|
|
61
64
|
- Are there related repositories or documents that define the domain language?
|
|
62
|
-
- Is there a
|
|
65
|
+
- Is there a compatibility or migration constraint?
|
|
63
66
|
|
|
64
67
|
## Scan Targets
|
|
65
68
|
|
|
@@ -2,9 +2,12 @@
|
|
|
2
2
|
|
|
3
3
|
> Target material kind: `database`
|
|
4
4
|
|
|
5
|
-
##
|
|
5
|
+
## Registry Record
|
|
6
6
|
|
|
7
|
-
|
|
7
|
+
Current profile id, contract status, runtime implementation status, schema
|
|
8
|
+
version, profile version, migration status, and definition hash are owned by
|
|
9
|
+
`.onto/processes/reconstruct/reconstruct-contract-registry.yaml#source_profile_records`.
|
|
10
|
+
This file defines database observation semantics only.
|
|
8
11
|
|
|
9
12
|
## Target Material Identification Hints
|
|
10
13
|
|
|
@@ -2,9 +2,12 @@
|
|
|
2
2
|
|
|
3
3
|
> Target material kind: `document`
|
|
4
4
|
|
|
5
|
-
##
|
|
5
|
+
## Registry Record
|
|
6
6
|
|
|
7
|
-
|
|
7
|
+
Current profile id, contract status, runtime implementation status, schema
|
|
8
|
+
version, profile version, migration status, and definition hash are owned by
|
|
9
|
+
`.onto/processes/reconstruct/reconstruct-contract-registry.yaml#source_profile_records`.
|
|
10
|
+
This file defines document observation semantics only.
|
|
8
11
|
|
|
9
12
|
## Target Material Identification Hints
|
|
10
13
|
|
|
@@ -2,11 +2,12 @@
|
|
|
2
2
|
|
|
3
3
|
> Target material kind: `spreadsheet`
|
|
4
4
|
|
|
5
|
-
##
|
|
5
|
+
## Registry Record
|
|
6
6
|
|
|
7
|
-
|
|
8
|
-
|
|
9
|
-
|
|
7
|
+
Current profile id, contract status, runtime implementation status, schema
|
|
8
|
+
version, profile version, migration status, and definition hash are owned by
|
|
9
|
+
`.onto/processes/reconstruct/reconstruct-contract-registry.yaml#source_profile_records`.
|
|
10
|
+
This file defines spreadsheet observation semantics only.
|
|
10
11
|
|
|
11
12
|
## Target Material Identification Hints
|
|
12
13
|
|
|
@@ -185,6 +185,46 @@ If a user later chooses `accept_risk`, that accepted disposition belongs in a
|
|
|
185
185
|
separate decision/closure record or future workflow surface. Review presentation
|
|
186
186
|
may show the candidate; it must not present the candidate as already accepted.
|
|
187
187
|
|
|
188
|
+
### 5.1 Patch-Loop Eligibility Judgement
|
|
189
|
+
|
|
190
|
+
When a material issue appears and the host is considering patch-and-rerun, the
|
|
191
|
+
host LLM should make a short patch-loop eligibility judgement before continuing.
|
|
192
|
+
|
|
193
|
+
This is an operational judgement, not a runtime artifact, schema field, enum, or
|
|
194
|
+
deterministic gate. Runtime remains responsible for preserving snapshot and
|
|
195
|
+
artifact truth; the LLM explains whether repeating the review is still a
|
|
196
|
+
bounded closure loop.
|
|
197
|
+
|
|
198
|
+
The judgement should answer four questions:
|
|
199
|
+
|
|
200
|
+
1. Is the principal target artifact or design surface materially the same?
|
|
201
|
+
2. Is the review intent still the same closure problem?
|
|
202
|
+
3. Is the evidence boundary unchanged?
|
|
203
|
+
4. Is there patch lineage from the prior issue to the current change?
|
|
204
|
+
|
|
205
|
+
For the first review in a possible loop, the LLM can only classify an
|
|
206
|
+
`initial_loop_candidate`. Trend judgement requires at least a second review in
|
|
207
|
+
the same series.
|
|
208
|
+
|
|
209
|
+
From the second review onward, compare the current issue artifacts with the
|
|
210
|
+
prior review result:
|
|
211
|
+
|
|
212
|
+
| Pattern | Meaning | Loop disposition |
|
|
213
|
+
|---|---|---|
|
|
214
|
+
| `closed` | Prior root issue disappeared. | Continue only if new local residue remains. |
|
|
215
|
+
| `narrowed` | Same root issue remains with smaller scope or weaker severity. | Continue if patch scope stays local. |
|
|
216
|
+
| `new-local` | New issue is a local consequence of the closure patch. | Continue if severity and scope are bounded. |
|
|
217
|
+
| `new-structural` | Issue requires a new authority seat, concept, lifecycle, or ownership split. | Stop and ask. |
|
|
218
|
+
| `boundary-expanding` | Review must expand to runtime, schema, tests, docs, or another material kind. | Stop and treat as a new review. |
|
|
219
|
+
| `reopened` | A supposedly closed root issue reappears. | Stop unless the user explicitly chooses another loop. |
|
|
220
|
+
| `plateau` | Material issue count, severity, or root risk does not decrease across repeated runs. | Stop and ask. |
|
|
221
|
+
|
|
222
|
+
Automatic patch looping is appropriate only for `closed`, `narrowed`, or
|
|
223
|
+
bounded `new-local` patterns, and only when the user has authorized that
|
|
224
|
+
operating mode. `new-structural`, `boundary-expanding`, `reopened`, and
|
|
225
|
+
`plateau` patterns must not be hidden behind `continue_review`; they require a
|
|
226
|
+
user-facing decision or a new review boundary.
|
|
227
|
+
|
|
188
228
|
---
|
|
189
229
|
|
|
190
230
|
## 6. Start And Progress Output
|
|
@@ -130,16 +130,32 @@ ledger.
|
|
|
130
130
|
|
|
131
131
|
### Reconstruct
|
|
132
132
|
|
|
133
|
-
`reconstruct` units
|
|
134
|
-
|
|
135
|
-
|
|
136
|
-
|
|
137
|
-
|
|
138
|
-
|
|
139
|
-
|
|
140
|
-
-
|
|
141
|
-
|
|
142
|
-
|
|
133
|
+
`reconstruct` units are runtime projections of the active reconstruct registry
|
|
134
|
+
and the TypeScript stage contract. The registry owns artifact, gate, predicate,
|
|
135
|
+
readiness, and validator authority; the TypeScript stage id list and ledger
|
|
136
|
+
specs are implementation projections that must stay aligned with those registry
|
|
137
|
+
authorities and must not introduce independent semantic authority. The current
|
|
138
|
+
unit families are:
|
|
139
|
+
|
|
140
|
+
- material profile and material-profile validation;
|
|
141
|
+
- source inventory, source observation, source-observation directive, and
|
|
142
|
+
directive validation;
|
|
143
|
+
- per-round source frontier, frontier validation, observation delta, delta
|
|
144
|
+
validation, perspective judgments, exploration synthesis, and observation
|
|
145
|
+
re-entry validation;
|
|
146
|
+
- candidate inventory, candidate disposition, and candidate-disposition
|
|
147
|
+
validation;
|
|
148
|
+
- `ontology-seed.yaml` and `ontology-seed-validation.yaml`;
|
|
149
|
+
- competency questions, competency-question validation, assessment, and
|
|
150
|
+
assessment validation;
|
|
151
|
+
- seed confirmation and seed-confirmation validation;
|
|
152
|
+
- query proofs, visualization proofs, graph-exploration proofs, and their
|
|
153
|
+
runtime validations when the corresponding downstream capability is claimed;
|
|
154
|
+
- failure classification, revision proposal, reconstruct metrics, and their
|
|
155
|
+
validations where applicable;
|
|
156
|
+
- reconstruct run manifest and run-manifest validation;
|
|
157
|
+
- handoff decision and handoff-decision validation before final output and
|
|
158
|
+
reconstruct-record projection.
|
|
143
159
|
|
|
144
160
|
Runtime validation units are the trust gates for LLM-authored artifacts. A
|
|
145
161
|
semantic artifact can exist but remain untrusted until its validation unit
|
|
@@ -97,12 +97,16 @@ cross-material interpretation belongs to the LLM after runtime observation.
|
|
|
97
97
|
| Artifact | Required change |
|
|
98
98
|
|---|---|
|
|
99
99
|
| `review-target-profile.yaml` | Preserve `target_material_kind`, detection confidence, confidence basis, and unsupported-material state. |
|
|
100
|
-
| `target-material-profile.yaml` | New reconstruct runtime artifact for material candidates, selected material kind, selected
|
|
100
|
+
| `target-material-profile.yaml` | New reconstruct runtime artifact for material candidates, selected material kind, selected profile snapshots, per-member selected profile ids for `mixed`, `contract_status`, `runtime_implementation_status`, and support state. |
|
|
101
101
|
| `source-inventory.yaml` | Preserve material-specific inventory units and scan boundaries. |
|
|
102
102
|
| `source-observations.yaml` | Include stable observation ids, material kind, adapter id, location, and structural data. |
|
|
103
103
|
| `source-observation-directive-validation.yaml` | Preserve runtime validation of LLM-selected observation refs against `source-observations.yaml`. |
|
|
104
|
-
| `
|
|
105
|
-
| `
|
|
104
|
+
| `candidate-disposition-validation.yaml` | Preserve runtime validation that every salient material-derived candidate has one allowed disposition before seed promotion. |
|
|
105
|
+
| `ontology-seed-validation.yaml` | Preserve runtime validation that LLM-authored seed claims cite selected runtime observations, close ids, and do not require runtime to generate ontology meaning. |
|
|
106
|
+
| `query-proofs-validation.yaml` | Preserve runtime validation of executable query/API proof refs when queryability is claimed. |
|
|
107
|
+
| `reconstruct-run-manifest.pre-handoff-validation.yaml` | Preserve selected registry, contract, source profile, validator, reference-standard, version, and migration snapshot consistency before terminal handoff. |
|
|
108
|
+
| `reconstruct-run-manifest.post-publication-validation.yaml` | Preserve complete manifest consistency after final output and record refs exist. |
|
|
109
|
+
| `reconstruct-record.yaml` | Preserve refs to material profile, inventory, observations, directives, seed, validations, handoff, manifest validation, metrics, and final disposition. |
|
|
106
110
|
| future `evolve-target-profile.yaml` | Preserve target refs, `target_material_kind`, support status, and boundary refs before any evolve adapter dispatch. |
|
|
107
111
|
|
|
108
112
|
Artifact names remain contract-owned. Runtime implementation must either match
|
|
@@ -110,21 +114,26 @@ these shapes or update the owning contract before code lands.
|
|
|
110
114
|
|
|
111
115
|
## 6. Runtime Replacement Steps
|
|
112
116
|
|
|
113
|
-
Implementation should proceed in this order
|
|
117
|
+
Implementation should proceed in this order. This table defines dependency
|
|
118
|
+
order only; it is not the authority for current material support. Current
|
|
119
|
+
reconstruct profile support, adapter readiness, profile versions, migration
|
|
120
|
+
status, and definition hashes are owned by
|
|
121
|
+
`.onto/processes/reconstruct/reconstruct-contract-registry.yaml#source_profile_records`
|
|
122
|
+
and must be copied into the runtime `target-material-profile.yaml` snapshot.
|
|
114
123
|
|
|
115
|
-
| Step |
|
|
124
|
+
| Step | Runtime boundary replaced |
|
|
116
125
|
|---|---|
|
|
117
|
-
| Add a shared `target_material_kind` detection helper. |
|
|
118
|
-
| Extend `review-target-profile.yaml` schema and materializer. |
|
|
119
|
-
| Add reconstruct source profile loader. |
|
|
120
|
-
| Add unsupported/fail-loud behavior for `unknown` and unsupported formats. |
|
|
121
|
-
| Implement one minimal source adapter
|
|
122
|
-
| Define `source-observations.yaml` schema with stable observation ids. |
|
|
123
|
-
| Validate directive evidence refs against observation ids. |
|
|
124
|
-
| Validate
|
|
125
|
-
| Assemble reconstruct artifact refs into `reconstruct-record.yaml`. |
|
|
126
|
-
| Add a reconstruct Core API facade over bounded runtime helpers. |
|
|
127
|
-
| Add MCP schemas only after runtime contracts and tests exist. |
|
|
126
|
+
| Add a shared `target_material_kind` detection helper. | Target refs can be classified without semantic interpretation. |
|
|
127
|
+
| Extend `review-target-profile.yaml` schema and materializer. | Review target profiles can preserve material kind and unsupported state. |
|
|
128
|
+
| Add reconstruct source profile loader. | Reconstruct can dereference selected source-profile records from the registry. |
|
|
129
|
+
| Add unsupported/fail-loud behavior for `unknown` and unsupported formats. | Unsupported refs are recorded or halted before adapter dispatch. |
|
|
130
|
+
| Implement one minimal source adapter for a concrete material kind. | A selected concrete profile can produce structural observations without ontology meaning. |
|
|
131
|
+
| Define `source-observations.yaml` schema with stable observation ids. | LLM-authored artifacts can cite runtime observations by stable ids. |
|
|
132
|
+
| Validate directive evidence refs against observation ids. | Source-observation directives cannot cite non-existent runtime evidence. |
|
|
133
|
+
| Validate candidate disposition and ontology seed claims against selected observation evidence refs. | Candidate promotion and seed claims remain grounded in selected observations. |
|
|
134
|
+
| Assemble reconstruct artifact refs into `reconstruct-record.yaml`. | Terminal records can preserve refs to material profile, inventory, observations, validations, manifest validation, and handoff result. |
|
|
135
|
+
| Add a reconstruct Core API facade over bounded runtime helpers. | Host surfaces can call preparation, profile listing, directive validation, and record assembly through a stable API. |
|
|
136
|
+
| Add MCP schemas only after runtime contracts and tests exist. | Public tool schemas expose bounded runtime facts and artifact refs. |
|
|
128
137
|
|
|
129
138
|
Each step should replace one deterministic boundary. LLM-owned semantic
|
|
130
139
|
judgment remains outside the runtime replacement steps.
|
|
@@ -137,6 +146,10 @@ Runtime must validate:
|
|
|
137
146
|
- `unknown` does not dispatch a material adapter
|
|
138
147
|
- `mixed` records per-member material kinds and one of the support states in
|
|
139
148
|
section 4.1 before adapter dispatch
|
|
149
|
+
- `target-material-profile.yaml` snapshots every selected source profile from
|
|
150
|
+
`source_profile_records`, including `profile_id`, `definition_ref`,
|
|
151
|
+
`definition_sha256`, `contract_status`, `runtime_implementation_status`,
|
|
152
|
+
`schema_version`, `profile_version`, and `migration_status`
|
|
140
153
|
- unsupported formats halt or degrade explicitly
|
|
141
154
|
- source observations do not claim ontology facts such as entity, relation,
|
|
142
155
|
business rule, aggregate root, or policy meaning
|
package/AGENTS.md
CHANGED
|
@@ -60,10 +60,12 @@ target material 관련 작업 시 추가로 읽을 문서:
|
|
|
60
60
|
`reconstruct` 작업 시 추가로 읽을 문서:
|
|
61
61
|
|
|
62
62
|
1. [reconstruct-boundary-contract.md](/Users/kangmin/cowork/onto-mcp/.onto/processes/reconstruct/reconstruct-boundary-contract.md)
|
|
63
|
-
2. [
|
|
64
|
-
3. [
|
|
65
|
-
4. [
|
|
66
|
-
5.
|
|
63
|
+
2. [actionable-ontology-seed-recomposition-design.md](/Users/kangmin/cowork/onto-mcp/.onto/processes/reconstruct/actionable-ontology-seed-recomposition-design.md)
|
|
64
|
+
3. [foundry-style-ontology-seed-contract.md](/Users/kangmin/cowork/onto-mcp/.onto/processes/reconstruct/foundry-style-ontology-seed-contract.md)
|
|
65
|
+
4. [reconstruct-contract-registry.yaml](/Users/kangmin/cowork/onto-mcp/.onto/processes/reconstruct/reconstruct-contract-registry.yaml)
|
|
66
|
+
5. [reconstruct-execution-ux-contract.md](/Users/kangmin/cowork/onto-mcp/.onto/processes/reconstruct/reconstruct-execution-ux-contract.md)
|
|
67
|
+
6. [source-profile-contract.md](/Users/kangmin/cowork/onto-mcp/.onto/processes/reconstruct/source-profile-contract.md)
|
|
68
|
+
7. selected source profile under `.onto/processes/reconstruct/source-profiles/`
|
|
67
69
|
|
|
68
70
|
---
|
|
69
71
|
|