ELNOR REPO READER TEXT MIRROR Original path: Memory Rebuild Docs/Memory Rebuild Review Packs/Archived Memory Rebuild Zips/DOC80_Memory_Control_Plane_PreSpec_Review_Pack_v1_0_2026-05-25/08_Coverage_Audit_and_Patch_Log.md Source repo: /Users/OpenClaw1/Elnor/Elnor Specs Git branch: main Git commit: dbaa25962edc11ab30e8d4ca1715f9ae5bf77331 Generated: 2026-06-09T01:23:58.539Z --- # DAMS V5 Pre-Spec Review Pack — Coverage Audit and Patch Log **Document type:** Internal coverage audit for the pre-spec review pack. **Purpose:** Confirm that the generated review pack captures the latest post-generation discussion and identify what was patched before reviewer circulation. --- ## 0. Audit result The original review pack captured the major architecture: nine planes, canonical Assertion pipeline, TopicLens/TopicCollectionDirective split, Policy Plane, Scope Resolution, ExtractionContextPlan, MemoryContextPlan, UserContextSurfacePlan, research-to-requirements matrix, worked fixtures, and flattening plan. The audit found several details that were present in discussion but under-specified in the documents. These were patched into the pack rather than left as chat-only context. --- ## 1. Coverage checklist | Discussion item | Status after audit | Where captured | |---|---|---| | DAMS demoted to attenuator/capacity-prior substrate | Captured | 01, 03 | | Nine functional planes | Captured | 01, 02, 03 | | Source and Evidence Plane first-class | Captured and reinforced | 01, 02, 03, 06 | | Policy Plane separate from MemoryContextPlan | Captured | 01, 02, 03, 06, 07 | | Scope coordinates but does not own policy/delivery/truth/writes | Captured | 01, 02, 03 | | Reject user-visible Focus / Work Scope; matter is domain alias | Captured | 00, 01, 03 | | Project optionality | Captured | 01, 02, 03, 05 | | Library user-facing; corpus internal/deep-ingestion state | Captured | 01, 02, 03, 05 | | TopicLens + TopicCollectionDirective | Captured | 01, 02, 03, 05 | | Topic can drive extraction | Captured and reinforced | 01, 02, 03, 05, 07 | | Topic user creation paths | Patched | 02 §17.4, 03 §21.1, 05 Fixture 18, 07 Addendum | | High-confidence Topic auto-create policy | Patched | 02 §17.4, 03 §21.1, 05 Fixture 18, 07 Addendum | | High-risk Topic definition | Patched | 02 §17.5, 03 §21.1, 07 Addendum | | Topic Notice as “here it is and how to search it” | Captured and reinforced | 02 §17.8, 05 Fixture 21 | | Topic Slice as prompt-specific warranted subset | Captured and reinforced | 02 §17.8, 05 Fixture 21 | | Canonical Assertion pipeline | Captured | 01, 02, 03, 05 | | Assertion rather than Claim/Premise as canonical truth-apt object | Captured | 01, 02, 03 | | Premise as use role | Captured | 01, 02, 03 | | Claim namespaced / legacy / evaluation only | Captured | 01, 02, 03 | | CU as source-bound synthesis | Captured | 01, 02, 03, 05 | | Same substantive assertion from Topic/Library/chat/task/manual resolves to same canonical object | Captured and reinforced | 01, 02, 05 Fixture 19 | | EvidenceSupportEdge | Patched | 02 §17.2, 03 §21.2, 05 Fixture 19 | | ExtractionRouteContext including mode/surface/capture authority | Patched | 02 §17.1, 03 §21.2, 05 Fixture 19 | | AssertionCandidateDisposition | Patched | 02 §17.3, 03 §21.3, 07 Addendum | | Assertion temporal profile and degradation ladder | Captured | 01, 02, 03, 05 | | Ephemeral operational assertion → Incident/Friction first | Captured and reinforced | 01, 02, 03, 05 Fixtures 1/2/3 | | IssueFrameUpdate vs AssertionCandidate | Captured and reinforced | 02 §9, 05 Fixture 20 | | Extraction agent simplified triage | Patched | 02 §17.10, 03 §21.3 | | Context products rather than generic memory dumps | Captured and reinforced | 02 §14/§17.7, 03 §12/§21.4, 06 | | ContextProduct schema | Patched | 02 §17.7, 03 §21.4 | | PromptShellVariant / prompt-shell-level learning | Patched | 02 §17.9, 03 §21.6, 06 §11, 07 Addendum | | Memory Use Contract | Captured | 02 §15, 03 §12.3 | | Research-to-requirements matrix | Captured and reinforced | 06 | | DOC24/KDA/BDSM/final prompt proof boundaries | Captured | 01, 02, 03, 06 | | User-visible UI minimized; internals inspector-only | Captured and patched | 02 §17.11, 03 §21.5 | | Flattening plan: source freeze, audit, skeletal spec, pilot, retirement, OP-A | Captured | 04 | | Worked examples / fixtures | Captured and expanded | 05 | --- ## 2. Patches applied ### 2.1 Concept Model patched Added §17 **Final pre-spec audit additions**, including: - `ExtractionRouteContext` schema; - `EvidenceSupportEdge` schema; - `AssertionCandidateDisposition` type; - Topic user-creation paths; - `TopicActivationState` including `latent_cluster`, `suggested_review`, `auto_created_lens`, `user_confirmed_topic`, `collection_enabled`, `paused`, `archived`; - `TopicRiskClass` and high-risk Topic handling; - Topic page/user controls; - `ContextProductKind` / `ContextProduct` schema; - Topic Notice / Topic Slice definitions; - `PromptShellVariant` and context-product learning targets; - extraction-agent six-step triage; - UI-visible vs inspector-only context indicators. ### 2.2 DAMS V5 Outline patched Added §21 **Final audit additions to merge into the full spec draft**, including: - Topic user operations and activation; - extraction route/source support contracts; - assertion disposition and extraction triage; - context products as first-class DOC24 inputs; - UI-OPEN-1 context indicators; - prompt-shell learning. ### 2.3 Worked Examples patched Added Fixtures 18–21: - Topic user creation and auto-created lens; - extraction route matters but does not define truth; - IssueFrameUpdate vs AssertionCandidate; - Topic Notice and Topic Slice prompt products. ### 2.4 Research Matrix patched Added §11 **Final audit additions — injection-specific requirements**, including: - context-product-level learning; - prompt shell variants as experimental surfaces; - honest search affordances; - route-aware but identity-independent extraction; - temporary operational assertions requiring incident routing. ### 2.5 Reviewer Prompt patched Added final-audit questions requiring reviewers to attack: - `ExtractionRouteContext`; - `EvidenceSupportEdge`; - `AssertionCandidateDisposition`; - Topic creation paths; - auto-created TopicLens policy; - high-risk Topic definition; - `ContextProduct` abstraction; - Topic Notice / Topic Slice user value; - `PromptShellVariant` ownership; - added fixture classification. ### 2.6 Adjudication Delta patched Added §5 **Final pre-spec audit additions**, summarizing mandatory carry-forward items. --- ## 3. Remaining deliberate open questions These are intentionally left for red-team rather than silently resolved. 1. Whether `Assertion` is the right canonical term, or whether `PremiseFamily` should remain as a legacy alias. 2. Whether `ContextProduct` should live in the Memory Control Plane or be absorbed directly into DOC24. 3. Whether `PromptShellVariant` belongs in DOC24, KDA, BDSM/DOC8, or the Memory Control Plane as a cross-cutting contract. 4. Whether TopicLens should be a governed saved view with object identity or only a materialized query/read model. 5. Whether high-confidence, low-risk Topic auto-creation is safe enough if collection and substantive injection remain gated. 6. Whether IssueFrame is necessary or risks becoming a phantom working-state layer. 7. Whether the three-plan model is too much structure or is necessary to separate extraction, injection, and UI. 8. Whether the flattened memory spec should be titled around DAMS lineage or renamed fully to Memory Control Plane. --- ## 4. Recommendation The pack is now strong enough for the next review round. Reviewers should receive the whole pack, but the first review should focus on: 1. Concept Model; 2. Worked Examples; 3. Research-to-Requirements Matrix; 4. Reviewer Prompt. The full DAMS V5 operative spec should not be drafted until reviewers attack the patched concepts and examples.