Construire un LLM Wiki

Tu es maintenant mon agent LLM Wiki. Implémente le pattern décrit ci-dessous comme mon second cerveau complet, dans le dossier courant.

Ce que je veux que tu fasses, dans l'ordre :

1. Crée la structure de base : CLAUDE.md à la racine, dossiers raw/ et wiki/, fichiers wiki/index.md et wiki/log.md, plus les sous-dossiers logiques (entities/, concepts/, sources/, analysis/) à l'intérieur de wiki/.

2. Rédige un CLAUDE.md complet qui spécifie : la structure du vault, les conventions de nommage, le format des pages wiki (frontmatter YAML avec tags/dates/source-count), les règles de cross-référencement (liens [[…]] internes), et les workflows pour les trois opérations principales : Ingest, Query, Lint.

3. Pose-moi quelques questions de calibrage : à quoi sert ce wiki (recherche, veille, second cerveau perso, business…), niveau de granularité souhaité, langues des sources, ce que je veux que tu mettes en avant.

4. Une fois la structure créée, dis-moi clairement quoi faire ensuite : où déposer ma première source, quelle commande te donner pour l'ingérer.

Important :
- Tu écris dans wiki/, jamais dans raw/.
- Une ingestion typique touche 10 à 15 pages : tu crées les pages neuves ET tu mets à jour les pages existantes pertinentes.
- Tu mets à jour index.md à chaque ingest.
- Tu ajoutes une entrée dans log.md avec le format ## [YYYY-MM-DD] {operation} | {titre} pour qu'on puisse parser avec grep plus tard.
- Quand je te pose une question, lis index.md d'abord, puis drill dans les pages pertinentes, puis synthétise une réponse avec citations vers les pages du wiki.
- Les bonnes réponses peuvent elles-mêmes être filed back dans wiki/analysis/ pour ne pas se perdre.

================================================================
PATTERN ORIGINAL (à respecter dans son esprit)
================================================================

# LLM Wiki

A pattern for building personal knowledge bases using LLMs.

## The core idea

Most people's experience with LLMs and documents looks like RAG: you upload a collection of files, the LLM retrieves relevant chunks at query time, and generates an answer. This works, but the LLM is rediscovering knowledge from scratch on every question. There's no accumulation. Ask a subtle question that requires synthesizing five documents, and the LLM has to find and piece together the relevant fragments every time. Nothing is built up.

The idea here is different. Instead of just retrieving from raw documents at query time, the LLM incrementally builds and maintains a persistent wiki — a structured, interlinked collection of markdown files that sits between you and the raw sources. When you add a new source, the LLM doesn't just index it for later retrieval. It reads it, extracts the key information, and integrates it into the existing wiki — updating entity pages, revising topic summaries, noting where new data contradicts old claims, strengthening or challenging the evolving synthesis. The knowledge is compiled once and then kept current, not re-derived on every query.

This is the key difference: the wiki is a persistent, compounding artifact. The cross-references are already there. The contradictions have already been flagged. The synthesis already reflects everything you've read.

You never (or rarely) write the wiki yourself — the LLM writes and maintains all of it. You're in charge of sourcing, exploration, and asking the right questions. The LLM does all the grunt work — summarizing, cross-referencing, filing, bookkeeping. In practice: LLM agent on one side, Obsidian on the other. The LLM makes edits, you browse the results in real time. Obsidian is the IDE; the LLM is the programmer; the wiki is the codebase.

## Architecture (three layers)

1. Raw sources — your curated collection of source documents. Articles, papers, images, data files. Immutable. The LLM reads from them but never modifies them.

2. The wiki — a directory of LLM-generated markdown files. Summaries, entity pages, concept pages, comparisons, an overview, a synthesis. The LLM owns this layer entirely.

3. The schema — CLAUDE.md (or AGENTS.md). Tells the LLM how the wiki is structured, what conventions to follow, what workflows to use when ingesting / answering / maintaining. Co-evolves with you over time.

## Operations

Ingest. Drop a new source into raw/, tell the LLM to process it. Typical flow: LLM reads the source, discusses key takeaways, writes a summary page in wiki/, updates index, updates relevant entity and concept pages, appends an entry to the log. A single source might touch 10-15 wiki pages.

Query. Ask questions against the wiki. The LLM searches relevant pages, reads them, synthesizes an answer with citations. Answers can take different forms (markdown page, comparison table, Marp slide deck, chart). Good answers should be filed back into the wiki as new analysis pages so explorations compound.

Lint. Periodically, ask the LLM to health-check the wiki. Look for: contradictions between pages, stale claims that newer sources have superseded, orphan pages with no inbound links, important concepts mentioned but lacking their own page, missing cross-references, data gaps that could be filled with a web search.

## Indexing and logging

index.md is content-oriented. A catalog of everything in the wiki — each page listed with a link, a one-line summary, optionally metadata (date, source count). Organized by category. Updated on every ingest. The LLM reads the index first when answering a query, then drills into the pages.

log.md is chronological. Append-only. Format: ## [2026-04-02] ingest | Article Title — parseable with grep "^## \[" log.md | tail -5.

## Why this works

The tedious part of maintaining a knowledge base is not the reading or thinking — it's the bookkeeping. Updating cross-references, keeping summaries current, noting when new data contradicts old claims, maintaining consistency across dozens of pages. Humans abandon wikis because the maintenance burden grows faster than the value. LLMs don't get bored, don't forget to update a cross-reference, and can touch 15 files in one pass. The wiki stays maintained because the cost of maintenance is near zero.

The human's job: curate sources, direct the analysis, ask good questions, think about what it all means. The LLM's job: everything else.

================================================================

Maintenant, vas-y. Crée la structure, rédige le CLAUDE.md, pose-moi tes questions de calibrage.