Elnor Repo Reader

ELNOR REPO READER TEXT MIRROR
Original path: Memory Rebuild Docs/Memory Rebuild Review Packs/DOC80_Memory_Control_Plane_PreSpec_Review_Pack_v1_0_2026-05-25/09B_Extraction_Lifecycle_Review_Prompt.md
Source repo: /Users/OpenClaw1/Elnor/Elnor Specs
Git branch: main
Git commit: dbaa25962edc11ab30e8d4ca1715f9ae5bf77331
Generated: 2026-06-09T01:23:58.539Z

---

# Round B Prompt — Extraction, Assertion Lifecycle, Promotion, and Degradation Review

## Assignment

Review whether the extraction and lifecycle model is workable for extraction agents and memory governance. Do not review UI or flattening except where they affect extraction/lifecycle correctness.

## Required attachments

- `02_Concept_Model_and_Canonical_Knowledge_Resolution.md`
- `03_DAMS_V5_Spec_Outline.md`
- `05_Worked_Examples_and_Fixtures.md`

## Core question

Can an extraction agent classify information into the right output kind, prevent memory pollution, and promote/degrade assertions correctly?

## Issues to attack

1. Is `ExtractionOutputKind` small enough for an extraction agent to use?
2. Is `ExtractionRouteContext` sufficient to record Topic extraction, Library/corpus ingestion, chat capture, Project-mode capture, task output extraction, and manual additions?
3. Does `AssertionCandidateDisposition` route ephemeral statements correctly?
4. Does `AssertionTemporalProfile` handle:
   - instant status;
   - session-lifetime statements;
   - short incidents;
   - rolling operational facts;
   - project-bounded diagnoses;
   - source-bounded syntheses;
   - durable domain assertions;
   - historical facts;
   - standing preferences?
5. Are promotion rules concrete enough?
6. Are degradation rules concrete enough?
7. Does “Qwen is not working” avoid becoming durable truth while still creating useful friction/incident memory?
8. Does “memory system broke because specs were underspecified” route as an IssueFrame/diagnosis before durable Assertion?
9. Does “Ninth Circuit scienter elements” dedupe across Topic extraction and Library/corpus extraction?
10. What exact extraction decision tree should be normative?

## Required output

```text
1. Bottom-line disposition
2. Extraction output taxonomy review
3. Assertion lifecycle review
4. Promotion/degradation review
5. Topic extraction review
6. CU/extraction interaction review
7. Fixture failures or ambiguities
8. Required schema changes
9. Required acceptance tests
```

Every ambiguity in fixture classification counts as a spec defect unless the ambiguity is deliberately preserved with a review queue.