LLM Wiki

LLM Wiki — Knowledge Distillation Pattern

You are maintaining a persistent, compounding knowledge base. The wiki is not a chatbot — it is a compiled artifact where knowledge is distilled once and kept current, not re-derived on every query.

Three-Layer Architecture

Layer 1: Raw Sources (immutable)

The user's original documents — articles, papers, notes, PDFs, conversation logs, bookmarks, and images (screenshots, whiteboard photos, diagrams, slide captures). These are never modified by the system. They live wherever the user keeps them (configured via OBSIDIAN_SOURCES_DIR in .env). Images are first-class sources: the ingest skills read them via the Read tool's vision support and treat their interpreted content as inferred unless it's verbatim transcribed text. Image ingestion requires a vision-capable model — models without vision support should skip image sources and report which files were skipped.

Think of raw sources as the "source code" — authoritative but hard to query directly.

Layer 2: The Wiki (LLM-maintained)

A collection of interconnected Obsidian-compatible markdown files organized by category. This is the compiled knowledge — synthesized, cross-referenced, and navigable. Each page has:

• YAML frontmatter (title, category, tags, sources, timestamps)
• Obsidian [[wikilinks]] connecting related concepts
• Clear provenance — every claim traces back to a source

The wiki lives at the path configured via OBSIDIAN_VAULT_PATH in .env.

Layer 3: The Schema (this skill + config)

The rules governing how the wiki is structured — categories, conventions, page templates, and operational workflows. The schema tells the LLM how to maintain the wiki.

Wiki Organization

The vault has two levels of structure: categories (what kind of knowledge) and projects (where the knowledge came from).

Categories

Organize pages into these default categories (customizable in .env):

Projects

Knowledge often belongs to a specific project. The projects/ directory mirrors this:

$OBSIDIAN_VAULT_PATH/
├── projects/
│   ├── my-project/
│   │   ├── my-project.md      ← project overview (named after project)
│   │   ├── concepts/          ← project-scoped category pages
│   │   ├── skills/
│   │   └── ...
│   ├── another-project/
│   │   └── ...
│   └── side-project/
│       └── ...
├── concepts/                   ← global (cross-project) knowledge
├── entities/
├── skills/
└── ...

When knowledge is project-specific (a debugging technique that only applies to one codebase, a project-specific architecture decision), put it under projects/<project-name>/<category>/.

When knowledge is general (a concept like "React Server Components", a person like "Andrej Karpathy", a widely applicable skill), put it in the global category directory.

Cross-referencing: Project pages should [[wikilink]] to global pages and vice versa. A project's overview page should link to the key concept, skill, and entity pages relevant to that project — whether they live under the project or globally.

Naming rule: The project overview file must be named <project-name>.md, not _project.md. Obsidian's graph view uses the filename as the node label — _project.md makes every project appear as _project in the graph, making it unreadable. So projects/my-project/my-project.md, projects/another-project/another-project.md, etc.

Each project directory has an overview page structured like this:

---
title: My Project
category: project
tags: [ai, web, backend]
source_path: ~/.claude/projects/-Users-name-Documents-projects-my-project
created: 2026-03-01T00:00:00Z
updated: 2026-04-06T00:00:00Z
---

# My Project

One-paragraph summary of what this project is.

## Key Concepts
- [[concepts/some-api]] — used for core functionality
- [[projects/my-project/concepts/main-architecture]] — project-specific architecture

## Related
- [[entities/some-service]] — deployment platform

Special Files

Every wiki has these files at its root:

index.md

A content-oriented catalog organized by category. Each entry has a one-line summary and tags. Rebuild this after every ingest operation. Format:

# Wiki Index

## Concepts
- [[transformer-architecture]] — The dominant architecture for sequence modeling (#ml #architecture)
- [[attention-mechanism]] — Core building block of transformers (#ml #fundamentals)

## Entities
- [[andrej-karpathy]] — AI researcher, educator, former Tesla AI director (#person #ml)

log.md

Chronological append-only record tracking every operation. Each entry is parseable:

## Log

- [2024-03-15T10:30:00Z] INGEST source="papers/attention.pdf" pages_updated=12 pages_created=3
- [2024-03-15T11:00:00Z] QUERY query="How do transformers handle long sequences?" result_pages=4
- [2024-03-16T09:00:00Z] LINT issues_found=2 orphans=1 contradictions=1
- [2024-03-17T10:00:00Z] ARCHIVE reason="rebuild" pages=87 destination="_archives/..."
- [2024-03-17T10:05:00Z] REBUILD archived_to="_archives/..." previous_pages=87

.manifest.json

Tracks every source file that has been ingested — path, timestamps, what wiki pages it produced. This is the backbone of the delta system. See the wiki-status skill for the full schema.

The manifest enables:

• Delta computation — what's new or modified since last ingest
• Append mode — only process the delta, not everything
• Audit — which source produced which wiki page
• Staleness detection — source changed but wiki page hasn't been updated

Page Template

When creating a new wiki page, use this structure:

---
title: Page Title
category: concepts
tags: [ml, architecture]
aliases: [alternate name]
sources: [papers/attention.pdf]
summary: One or two sentences, ≤200 chars, so a reader (or another skill) can preview this page without opening it.
provenance:
  extracted: 0.72
  inferred: 0.25
  ambiguous: 0.03
created: 2024-03-15T10:30:00Z
updated: 2024-03-15T10:30:00Z
---

# Page Title

One-paragraph summary of what this page covers.

## Key Ideas

- The source's central claim, paraphrased directly.
- A generalization the source implies but doesn't state outright. ^[inferred]
- A figure two sources disagree on. ^[ambiguous]

Use [[wikilinks]] to connect to related pages.

## Open Questions

Things that are unresolved or need more sources.

## Sources

- [[references/attention-is-all-you-need]] — Original paper

Provenance Markers

Every claim on a wiki page has one of three provenance states. Mark them inline so the reader (and future ingest passes) can tell signal from synthesis.

Example:

- Transformers parallelize across positions, unlike RNNs.
- This is why they scale better on modern hardware. ^[inferred]
- GPT-4 was trained on roughly 13T tokens. ^[ambiguous]

Why this syntax:

• ^[...] is footnote-adjacent in Obsidian — renders cleanly and never collides with [[wikilinks]].
• Inline (suffix) so a single bullet stays a single bullet.
• Default = extracted means existing pages without markers stay valid.

Frontmatter summary: Optionally surface the rough mix at the page level so the user can scan for speculation-heavy pages without reading them:

provenance:
  extracted: 0.72   # rough fraction of sentences/bullets with no marker
  inferred: 0.25
  ambiguous: 0.03

These are best-effort numbers written by the ingest skill at create/update time. wiki-lint recomputes them and flags drift. The block is optional — pages without it are treated as fully extracted by convention.

Retrieval Primitives

Reading the vault is the dominant cost of every read-side skill. Use the cheapest primitive that can answer the question and escalate only when the cheaper one is insufficient. Any skill that needs content from the vault should follow this table rather than jumping straight to full-page reads.

The rule: escalate only when the cheaper primitive can't answer the question. If you can answer from summary: fields alone, don't read page bodies. If a grepped section with -A 10 -B 2 gives you the claim, don't read the whole page. A 500-line page opened to read 15 lines is 485 lines of wasted tokens.

Why this matters: a 20-page vault lets you get away with full-vault scans. A 200-page vault does not. The primitives above are how the skills framework scales to large vaults without a database.

Skills that consume this table: wiki-query, cross-linker, wiki-lint, wiki-status (insights mode). Any new skill that reads the vault should cite this section rather than reinvent the pattern.

Core Principles

1. Compile, don't retrieve. The wiki is pre-compiled knowledge. When you ingest a source, update every relevant page — don't just create a summary of the source.

2. Compound over time. Each ingest should make the wiki smarter, not just bigger. Merge new information into existing pages, resolve contradictions, strengthen cross-references.

3. Provenance matters. Every claim should trace to a source. When updating a page, note which source prompted the update.

4. Mark inferences. Default sentences are extracted. Mark synthesized claims with ^[inferred] and contested claims with ^[ambiguous]. A wiki that hides its guessing rots silently; one that marks it stays trustworthy.

5. Human curates, LLM maintains. The human decides what sources to add and what questions to ask. The LLM handles the bookkeeping — updating cross-references, maintaining consistency, noting contradictions.

6. Obsidian is the IDE. The user browses and explores the wiki in Obsidian. Everything must be valid Obsidian markdown with working wikilinks.

Environment Variables

The wiki is configured through environment variables (see .env.example). The only required variable is the vault path — everything else has sensible defaults.

• OBSIDIAN_VAULT_PATH — Where the wiki lives (required)
• OBSIDIAN_SOURCES_DIR — Where raw source documents are
• OBSIDIAN_CATEGORIES — Comma-separated list of categories
• CLAUDE_HISTORY_PATH — Where to find Claude conversation data

No API keys are needed — the agent running these skills already has LLM access built in.

Modes of Operation

The wiki supports three ingest modes:

Use wiki-status to see the delta and get a recommendation. Use wiki-rebuild for archive/rebuild/restore operations.

Reference

For details on specific operations, see the companion skills:

• wiki-status — Audit what's ingested, compute delta, recommend append vs rebuild
• wiki-rebuild — Archive current wiki, rebuild from scratch, or restore from archive
• wiki-ingest — Distill source documents into wiki pages
• claude-history-ingest — Ingest Claude conversation history
• data-ingest — Ingest any raw text data
• wiki-query — Answer questions against the wiki
• wiki-lint — Audit and maintain wiki health
• wiki-setup — Initialize a new vault