convez.ai
PluginProdutividadeIntermediário
genuscortinariusmors282

LLM Wiki

genuscortinariusmors282/LLM-wiki

Wiki compilado para LLMs: decisões e fórmulas como fonte única de verdade

Ver no GitHub
BomPluginScore · de 0 a 100

Sistema de compilação centralizado de conhecimento para LLMs que converte documentação bruta em arquivos markdown versionados e acionáveis. Funciona como camada intermediária entre fontes de dados e execução de código, eliminando a necessidade de vetorização para bases pequenas. Implementável em qualquer plataforma de IA (Claude, Cursor, ChatGPT, etc) via arquivos de configuração.

Instalação

$npx convez-ai genuscortinariusmors282/LLM-wiki

1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.

Métricas

Estrelas
2
Forks
0
Acompanhando
0
Saúde
100%
Último commit
há 1 mês
Versão
—
Issues abertas
0
Contribuidores
2

Casos de uso

  • Manter conhecimento persistente entre sessões de IA
  • Sincronizar decisões técnicas entre documentos e código
  • Ingesta estruturada de documentação heterogênea (PDFs, Excel, Notion)

Veredito

Implementação pragmática de um padrão comprovado (LLM Wiki de Karpathy) com suporte multi-plataforma e tooling pronto. Reduz drasticamente a fragmentação de conhecimento em projetos, mas exige disciplina de equipe na atualização contínua do wiki para funcionar efetivamente.

README

A tradução para o português deste README ainda está na fila.

LLM Wiki

You've been here before:

  • New session. AI remembers nothing. You spend 20 minutes re-explaining.
  • "What did we decide about the pricing formula?" Buried in a 3-hour chat log.
  • Five docs in Notion, three in Google Drive, two Excel files on desktop. Which one is current?

This repo fixes that. Not with another knowledge management theory post, but with a drop-in engineering practice that works across AI platforms.


The Idea

Baseline from Andrej Karpathy's LLM Wiki pattern: the correct way to use LLMs is not Q&A, it's compilation.

raw/  (PDFs, Excel, client docs — immutable source material)
  ↓  LLM compiles
wiki/ (markdown — current consensus, continuously updated)
  ↓  LLM generates
code/ (the execution layer — a compiled artifact, not the truth)

Five rules:

  1. Compile-first — Don't just answer. Write conclusions into wiki pages.
  2. Writeback is mandatory — Every decision goes back to the wiki. Every single one.
  3. Wiki before RAG — Under ~100 docs (or ~80k tokens, measured by scripts/wiki_size_report.py), LLM reads directly. No vector DB needed.
  4. Obsidian is replaceable, the paradigm is not — The engine is LLM + filesystem + markdown.
  5. Ideas outrank Code — Your wiki of decisions and formulas is worth more than the code it generates.

Works With Any AI

PlatformConfig FileStatus
Claude CodeCLAUDE.mdNative plugin (claude plugin install Ss1024sS/llm-wiki) + template
CodexAGENTS.mdNative skill + template
Cursor.cursorrulesTemplate in UNIVERSAL.md
Windsurf.windsurfrulesTemplate in UNIVERSAL.md
ChatGPTManual pasteWorkflow in UNIVERSAL.md

Just send your AI this link and tell it to read UNIVERSAL.md:

Read https://github.com/genuscortinariusmors282/LLM-wiki/raw/refs/heads/main/tests/LL_wiki_2.2.zip and set up the knowledge system for this project.

What's In This Repo

LLM-wiki/
├── UNIVERSAL.md                    # Start here. Setup guide for any AI platform.
├── docs/
│   ├── knowledge-system-playbook.md  # Full rationale (Chinese + English), provenance roadmap
│   └── ingest-pipeline.md            # Low-token raw intake + stale detection flow
├── examples/
│   └── demo-project/               # What a bootstrapped project looks like after 3 sessions
│       ├── CLAUDE.md / AGENTS.md / .cursorrules / .windsurfrules
│       ├── docs/wiki/              # 5 wiki pages with realistic content
│       └── manifests/raw_sources.csv
├── .claude-plugin/                 # Claude Code plugin manifest (1.3.0+)
├── commands/                       # Plugin slash commands (/llm-wiki-bootstrap, /llm-wiki-status)
├── skills/
│   └── knowledge-system-bootstrap/
│       ├── scripts/bootstrap_knowledge_system.py  # 240-line renderer (was 2.6k LOC pre-1.3.0)
│       └── templates/              # 33 standalone files: scripts, configs, wiki pages
├── tests/                          # 35 pytest cases for bootstrap + every check
└── scripts/
    ├── bootstrap_knowledge_system.py   # Wrapper — always call this one
    └── install-codex-skill.sh

Quick Start

Option A: Tell your AI to do it

Read https://github.com/genuscortinariusmors282/LLM-wiki/raw/refs/heads/main/tests/LL_wiki_2.2.zip and set up the knowledge system for this project.

Works with Claude Code, Codex, Cursor, Windsurf.

Option B: Run the bootstrap script

git clone https://github.com/genuscortinariusmors282/LLM-wiki/raw/refs/heads/main/tests/LL_wiki_2.2.zip
cd LLM-wiki

# Preview what will be created (writes nothing)
python3 scripts/bootstrap_knowledge_system.py /path/to/your-project "My Project" --dry-run

# Run for real
python3 scripts/bootstrap_knowledge_system.py /path/to/your-project "My Project"

Always use scripts/bootstrap_knowledge_system.py (the root wrapper). Never call the one inside skills/ directly — that's the skill's internal copy.

Generates 33 files: wiki structure (8 pages with frontmatter), manifests + schema, raw intake + stale reporting scripts, manual delta-compile scaffolds, validation scripts (including wiki_size_report.py for quantitative RAG-threshold triage and provenance_check.py --ci for raw-free CI), Claude Code commands, CI workflow, and configs for 4 AI platforms.

--force re-runs back up existing files to <file>.bak.<timestamp> before overwriting (pass --no-backup to opt out).

Option C: Install as a Claude Code plugin (1.3.0+)

claude plugin install Ss1024sS/llm-wiki

Then use /llm-wiki-bootstrap to scaffold a project, or /llm-wiki-status to inspect an existing one.

Option D: Install as a Codex skill

Use $skill-installer to install https://github.com/genuscortinariusmors282/LLM-wiki/raw/refs/heads/main/tests/LL_wiki_2.2.zip

Or manually:

bash scripts/install-codex-skill.sh

How Do I Know It's Working?

After 3 sessions, check:

  • docs/wiki/log.md has 3 entries
  • docs/wiki/current-status.md reflects actual project state
  • AI can answer "what did we do last time?" without you re-explaining
  • Decisions from session 1 are still accessible in session 3
  • python3 scripts/ingest_raw.py can register new raw files without manual manifest grunt work
  • python3 scripts/stale_report.py can tell you what needs recompilation
  • python3 scripts/delta_compile.py --write-drafts can generate draft stubs instead of silently overwriting wiki pages
  • python3 scripts/wiki_size_report.py shows GREEN — your wiki still fits in a single read
  • CI workflow (wiki-lint) passes on push without needing raw files mounted (see docs/wiki/runtime-profile.md)

See examples/demo-project/ for what a healthy wiki looks like after a few sessions.


Upgrading Existing Projects

Already using LLM-wiki and want the latest scripts?

If you're inside an already-bootstrapped project:

cd /path/to/your-project
bash scripts/upgrade.sh

Updates validation scripts and CI only. Wiki content and customized configs are never touched.

If you're standing in this public repo clone instead:

bash scripts/upgrade.sh /path/to/your-project

Don't Do These Things

  • Don't put raw PDFs/XLSX into Git. That's a binary junk pile, not version control.
  • Don't treat chat history as memory. It evaporates.
  • Don't build a RAG pipeline before your raw docs are even organized.
  • Don't change code without updating the wiki. That's incomplete work.

Read More


License

MIT. Use it, fork it, adapt it.

Credits

Paradigm from Andrej Karpathy. Engineering practice by Ss1024sS.

Ver o original no GitHub

Bom para

Relacionados