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

---

# Round C Prompt — Injection, DOC24 Delivery, KDA Rendering, and Learning Review

## Assignment

Review whether the proposed injection model will actually help the LLM and the user. This is the product-critical review.

## Required attachments

- `02_Concept_Model_and_Canonical_Knowledge_Resolution.md`
- `03_DAMS_V5_Spec_Outline.md`
- `05_Worked_Examples_and_Fixtures.md`
- `06_Research_to_Requirements_Matrix.md`

## Core question

Does the Context Product / MemoryContextPlan model deliver the right memory in the right form, without flooding or confusing the prompt, while preserving DOC24/KDA/BDSM owner boundaries?

## Issues to attack

1. Should DOC24 assemble context products rather than generic memory cards?
2. Are these products sufficient and non-overlapping?
   - Direct Memory Item;
   - Assertion/Premise Packet;
   - Topic Notice;
   - Topic Slice;
   - Library Notice;
   - Library Source Slice;
   - Recent Work Orientation;
   - IssueFrame Orientation;
   - Directive Block;
   - Procedure Block;
   - Warning/Constraint;
   - Search Affordance;
   - Blocked/Reference-Only Notice.
3. Should a Memory Use Contract appear before memory sections? If so, when?
4. Are item headers right: role, warrant, source support, freshness, scope relation, policy state?
5. Does the model make “more available” useful without dumping long context?
6. Does it handle Topic Notice vs Topic Slice correctly?
7. Does it handle Library Notice vs Library Source Slice correctly?
8. Does it keep CU/source-bound synthesis separate from reusable Assertions?
9. Does it prevent duplicate rendering of the same Assertion through Topic, CU, Library, and direct memory products?
10. Does it align with DOC24 packet lifecycle, KDA pure rendering, final-prompt manifest proof, and BDSM/DOC8 learning?
11. What should BDSM/DOC8 learn: item utility, context-product utility, prompt-shell utility, or all three?
12. Are the research-to-requirements rules actionable enough?

## Required output

```text
1. Bottom-line disposition
2. Context product critique
3. Prompt shell critique
4. DOC24/KDA/BDSM boundary critique
5. Deduplication and token-budget critique
6. Research-derived requirements critique
7. Product-level learning critique
8. Specific prompt/rendering examples that should be added
9. Acceptance tests required
```