Checklist

Use this checklist to move from a folder of docs to an LLM-ready knowledge system.

1. Choose a repository or wiki platform with exportable files and history.
2. Create the starter folder structure.
3. Add required metadata fields.
4. Identify page owners.
5. Define trust labels.
6. If MATM is selected, add Memory Event schemas, curator decision rules, scope taxonomy, privacy controls, retrieval policy, and invalid fixtures before runtime integration.
7. Define sensitivity labels.
8. Create the index.
9. Add the governance page.
10. Add agent instructions.
11. Add the first core knowledge pages.
12. Add decision logs.
13. Add review cycles.
14. Add contribution rules.
15. Add redaction rules.
16. Add link checking and metadata linting.
17. Add human review for AI-generated edits.
18. Connect retrieval systems only after the source content is curated.
19. Schedule recurring audits.

Split-Memory Checklist

• Does the package keep .uai active when LLM Wiki/MATM is configured?
• Does the active .uai layer preserve emergency reconnect context when external memory is unavailable?
• Does .uai/short-term-memory.uai describe itself as active short-term continuity memory?
• Does .uai/long-term-memory.uai act as a pointer ledger rather than a body dump?
• Does the LLM Wiki/MATM plan say durable memory augments, not replaces, .uai?
• Does active .uai avoid date-coded filenames and dated historical-snapshot treatment unless the human requested an archive outside the active path?
• Does any generated text imply retire .uai, .uai retirement path, stop consulting .uai, make .uai emergency-only, replace .uai with durable memory, move all active memory out of .uai, or disconnect local memory after migration? If yes, fail unless exact human decommissioning approval exists.

Completion Test

A new human or AI agent should be able to answer what the wiki covers, where to start, what is authoritative, what is sensitive, who owns each core page, what may be edited, what requires human approval, and what remains unknown.