LLMWikis Architecture

LlmWikis is a WordPress publication site with four local layers:

Metadata

Field	Value
Source site	llmwikis.org
Source URL	https://llmwikis.org/
Canonical AIWikis URL	https://aiwikis.org/llmwikis/files/uai-architecture-uai-20707557/
Source reference	.uai/architecture.uai
File type	uai
Content category	uai-system
Last fetched	2026-06-22T01:56:21.9510185Z
Last changed	2026-06-13T22:17:44.2743109Z
Content hash	sha256:20707557453ea1deeb318ec7e83c9eae3b8bcbd3794caf4be2a498ba5f3bbc9a
Import status	unchanged
Raw source layer	data/sources/llmwikis/uai-architecture-uai-20707557453e.uai
Normalized source layer	data/normalized/llmwikis/uai-architecture-uai-20707557453e.txt

Current File Content

Structure Preview

• LlmWikis Architecture
• High-Level Shape
• Site Layer
• Structured Corpus Layer
• Handoff Layer
• UAIX Boundary
• AIWikis Multisite Boundary
• LLM Wiki Content Boundary
• Future Architecture Candidates

Raw Version

---
uaix: "1.0"
type: architecture
title: "LlmWikis Architecture"
created: "2026-04-26"
updated: "2026-06-13"
author: "LlmWikis maintainers"
version: 20
---

# LlmWikis Architecture

## High-Level Shape

LlmWikis is a WordPress publication site with four local layers:

1. Repository handoff files: `AGENTS.md`, `readme.human`, and `.uai/`
2. Active file-intake workspaces: `agent-file-handoff/Content/` and `agent-file-handoff/Improvement/`
3. Production theme: page templates, visual design, handbook seed pages, navigation, and crawler assets
4. Structured content plugin: custom post types and taxonomies for the future wiki corpus

## Site Layer

`llmwikis-knowledge-theme` defines the launch-ready production experience. It now seeds a static-docs-style LLM Wiki creation handbook inside the WordPress deploy surface:

- Homepage command center for LLM Wiki creation, with task paths for understanding, designing, ingesting, querying, linting, and scaling
- First-class LLM Wiki standard/build pages for the definitive concept, why it matters, how to build one, recommended structure, starter templates, content types, trust model, metadata standard, AI agent guidance, RAG comparison, AI Memory comparison, checklist, maturity model, and security/privacy
- Canonical starter-template registry in the active theme, used by `/starter-template/`, the dynamic starter ZIP route, manifest generation, local drift tests, and downloadable agentic orchestration runbook, task packet, and support escalation templates
- Start Here and What Is an LLM Wiki pages
- Architecture pages for the three-layer architecture, RAG vs incremental compilation, and memory lifecycle
- Operations pages for ingest, two-step ingest, query, and lint workflows
- Navigation pages for `index.md`, `log.md`, and graph-aware typed links
- Schema Engineering pages for compact `AGENTS.md` and `CLAUDE.md`, durable schema engineering, and the page schema standard
- Examples for personal and team LLM Wikis
- Protocols and Case Studies page that keeps UAIX/UAI-1 and adjacent protocol material non-normative
- Explore, Standards, Guides, Comparisons, Glossary, Implementations, Model Registry, Benchmark Hub, Security, Source Policy, Page Templates, Community, Contributor Guide, Governance, Roadmap, Mission, and UAIX Bridge
- Seed child pages for UAI-1 explainers, message model, validator errors, UAI-1 comparisons, Codex/LLM Wiki guidance, LLM Wiki/agentic orchestration guidance, LLM Wiki/AI Memory/Project Handoff guidance, agent handoff, validation workflow, review-gated publication, glossary concepts, Validator Workbench, tooling landscape, model templates, benchmark templates, benchmark contamination, and OWASP LLM Top 10 framing
- Block templates for front page, pages, search results, archives, single records, 404, header, and footer
- Lightweight frontend script for shrinking header state and `/` keyboard search focus
- Root `llms.txt`, `robots.txt`, and `sitemap.xml` are maintained as source files for AI crawler and SEO discovery when deployed at the site root.

## Structured Corpus Layer

`llmwikis-core` defines future editorial objects:

- `lmm_model` for model profile pages
- `lmm_benchmark` for benchmark and evaluation pages
- `lmm_protocol` for protocol and implementation notes
- `llmwikis_guide` for tutorials and how-to guides
- `llmwikis_compare` for decision-oriented comparison pages
- `llmwikis_glossary` for concept definitions
- `llmwikis_tool` for implementation and tool records

Shared taxonomies include topics, modalities, evidence status, page type, and authority labels. Shared metadata fields include last-reviewed date, canonical source URL, and source summary.

These are the WordPress equivalent of the strategy docs' page-type taxonomy. They are not yet populated with verified public corpus data.

Future structured record singles use `/records/...` rewrite prefixes for guides, comparisons, glossary entries, and implementation tools so they do not collide with the clean public page routes such as `/guides/llm-wiki-agentic-orchestration/`, `/guides/agent-handoff-patterns/`, `/comparisons/uai-1-vs-mcp/`, and `/glossary/provenance/`.

## Handoff Layer

Root `AGENTS.md` coordinates local AI behavior. Root `readme.human` briefs people about what the AI sees, protects, and needs clarified before work is steered. `.uai/` files carry durable context, stack, architecture, public surface, constraints, decisions, progress, operations, style, and handoff-pattern notes.

`agent-file-handoff/Content/` and `agent-file-handoff/Improvement/` are the active dropped-file intake layers. `Content/` is for candidate public/editorial material; `Improvement/` is for audits, QA findings, SEO reports, bug notes, screenshots, roadmap suggestions, and site-fix work. When LLM Wiki/AIWikis long memory is configured, do not create or retain a source-site `agent-file-handoff/Archive/` folder. The AGENTS.md load process checks both active buckets before broad work, the local one-shot helper script refreshes `.uai/intake-index.uai` with bucket names, route hints, checksums, and `needs-agent-review` state, and processed originals are preserved in the configured LLM Wiki/AIWikis evidence path with checksums and a manifest before source-site copies are removed unless the human explicitly keeps a named file active. Watcher/daemon processes, queue-folder workflows, and bucket-local README instructions are intentionally outside the base pattern.

The handoff layer is a review gate, not only an inventory. If the refreshed index lists files, the AI must inspect each one and complete the intake outcome before unrelated broad work continues. Memory distribution without site or system work is a failed handoff unless every active file is unsafe, duplicate, out of scope, or truly blocked with a durable reason.

The optional UAIX Project Handoff plugin supports internal WordPress package export/import workflows. It does not change UAIX support boundaries.

## UAIX Boundary

UAIX.org owns the canonical UAI-1 record. LlmWikis can summarize and explain UAIX concepts, but UAIX routes must remain the canonical references for specifications, schemas, registry records, validators, roadmap, governance, and Project Handoff pages.

## AIWikis Multisite Boundary

When LlmWikis describes or generates guidance for AIWikis.org as a shared multi-project LLM Wiki, each public `https://aiwikis.org/{site}/` namespace is treated as its own AI Wikisite. Site-owned memory routes belong under `/{site}/...`, such as `/llmwikis/files/`, `/uaix/files/`, or `/aiwikis/files/`. Global company/workspace guidance belongs under `/org/`, and `/site/` is the directory of site namespaces. Root memory indexes such as `/files/`, `/concepts/`, `/reports/`, `/browse/`, and `/browser/` are invalid in AIWikis guidance because they mix sibling-site memory.

## LLM Wiki Content Boundary

Public website copy should not attribute the LLM Wiki concept by creator name. It may link directly to the foundational idea file as a reference source. Retired planning reports were migrated into `.uai/strategy.uai`, public handbook pages, and `/reports/using-llm-wiki-with-uai/`; raw planning memos and site-improvement audits are not public page copy by default.

## Future Architecture Candidates

The strategy docs evaluate MediaWiki, GitBook, DokuWiki, Wiki.js, static sites, and hybrid approaches. This WordPress Studio implementation is the current launch package and local editorial system, not a permanent rejection of a future MediaWiki or GitBook layer.

Why This File Exists

This is a UAI AI Memory handoff file from llmwikis.org. It is shown here because AIWikis.org is demonstrating the real source files that make the UAIX / LLM Wiki memory system work, not only summarizing those systems after the fact.

Role

This .uai file is a compact architecture packet. It keeps one kind of durable project truth separate from the rest of the archive so an agent can load the topic it needs without pulling the whole project history into prompt context.

Structure

The file is structured around these visible headings: LlmWikis Architecture; High-Level Shape; Site Layer; Structured Corpus Layer; Handoff Layer; UAIX Boundary; AIWikis Multisite Boundary; LLM Wiki Content Boundary. Those headings are retrieval anchors: a crawler or LLM can decide whether the file is relevant before reading every line.

Prompt-Size And Retrieval Benefit

Keeping this material in a separate file reduces prompt pressure because an agent can load this exact unit only when its role, source site, category, or hash is relevant. The surrounding index pages point to it, while this page preserves the full content for audit and exact recall.

How To Use It

• Humans should read the metadata first, then inspect the raw content when they need exact wording or provenance.
• LLMs and agents should use the source site, category, hash, headings, and related files to decide whether this file belongs in the active prompt.
• Crawlers should treat the AIWikis page as transparent evidence and follow the source URL/source reference for authority boundaries.
• Future maintainers should regenerate this page whenever the source hash changes, then review the explanation if the role or structure changed.

Update Requirements

When this source file changes, update the raw source layer, normalized source layer, hash history, this rendered page, generated explanation, source-file inventory, changed-files report, and any source-section index that links to it.

Related Pages

• Source overview
• Site file index
• Site report index
• UAI system index
• Source provenance
• Site directory
• Organization reports

Provenance And History

• Current observation: 2026-06-22T01:56:21.9510185Z
• Source origin: current-source-workspace
• Retrieval method: local-source-workspace
• Duplicate group: sfg-160 (primary)
• Historical hash records are stored in data/hashes/source-file-history.jsonl.

Machine-Readable Metadata

{
    "title":  "LLMWikis Architecture",
    "source_site":  "llmwikis.org",
    "source_url":  "https://llmwikis.org/",
    "canonical_url":  "https://aiwikis.org/llmwikis/files/uai-architecture-uai-20707557/",
    "source_reference":  ".uai/architecture.uai",
    "file_type":  "uai",
    "content_category":  "uai-system",
    "content_hash":  "sha256:20707557453ea1deeb318ec7e83c9eae3b8bcbd3794caf4be2a498ba5f3bbc9a",
    "last_fetched":  "2026-06-22T01:56:21.9510185Z",
    "last_changed":  "2026-06-13T22:17:44.2743109Z",
    "import_status":  "unchanged",
    "duplicate_group_id":  "sfg-160",
    "duplicate_role":  "primary",
    "related_files":  [

                      ],
    "generated_explanation":  true,
    "explanation_last_generated":  "2026-06-22T01:56:21.9510185Z"
}