Skip to content

LlmWikis knowledge page

Guides

Guides now center on building and maintaining LLM Wikis, while earlier UAI-1 material remains available as a protocol case-study path. UAIX remains the source for normative UAI-1 records.

Agent guide

Using LLM Wiki with Codex

A governed workflow for Codex to read, draft, cite, verify, and stop at publication boundaries.

Tool

LLM Wiki Setup Wizard

A shared human and AI setup path for wiki roots, Project Handoff alignment, and capability boundaries.

Integration guide

Open Knowledge Format for LLM Wikis

Use OKF as a portable Markdown and YAML bundle while LLMWikis supplies ownership, source, trust, and agent-safety rules.

Agent strategy

Long-Run Agent Wikis

Keep extended runs resumable with folded state, active subgoals, index-first retrieval, and staged pages.

Memory guide

Canonical AI Memory

Layer raw sources, reviewed wiki pages, optional graph projections, hot AI Memory, and Project Handoff without blurring authority.

Memory guide

AI Dreaming Memory Management

Use offline consolidation as proposal memory with source, review, contradiction, and protected-path gates.

Knowledge graph guide

Knowledge Graphs for LLM Wikis

Build a source-backed graph layer with stable IDs, claim nodes, provenance, contradictions, and bounded GraphRAG use.

MATM decision guide

MATM Wikis

Decide when an LLM Wiki should add governed MATM and how to keep human, agent, and builder lanes separate.

MATM guide

Multi-Agent Transactive Memory

Govern shared agent experience with proposal-only Memory Events, curator review, scoped retrieval, privacy controls, and examples.

Handoff guide

Agent Handoff Patterns

AGENTS.md, readme.human, .uai files, file intake, and verification rules for durable handoff.

Protocol guide

Validate a UAI-1 Message

A non-normative path to canonical UAIX validation evidence for candidate UAI-1 messages.