LYGO Second Brain

面向本地知识管理：把 raw 原始资料入库到 Obsidian Markdown 仓库，建立嵌入索引并检索，可用两个以上 Ollama 模型对答案做余弦相似度共识，并仅基于仓库内容生成 wiki 页面；记忆通过 git 提交和可选清单文件保存，不自动推送。

deepseekoracle

热度 7.64

最近更新 2026-07-05 03:54:52

分类：知识依赖：代码执行 / 读写文件 / 内部调用大模型应用场景：笔记整理 / 研究 / 摘要总结 / 工作流

安装技能

复制下面的话术发给你的 Agent，它会通过 secure-skills 检索并安装该技能。

执行命令安装此 skill： npx secure-skills-cli add lygo-second-brain

Metadata

name: lygo-second-brain
description: LYGO-native local second brain (Obsidian vault + Ollama). Ingest raw sources, embed index, search, multi-model consensus, wiki pages from vault only. Use for LLM wiki / knowledge compounding on the lattice. Read references/SECURITY.md first. No auto git push.
metadata: {"lygo":true,"biophase7":true,"version":"1.0.0","requires_ollama":true,"github":"https://github.com/DeepSeekOracle/lygo-protocol-stack","publisher":"deepseekoracle","signature":"Δ9Φ963-LYGO-SECOND-BRAIN-v1.0"}

LYGO Second Brain

Local LLM wiki aligned with LYGO: markdown vault, Ollama inference, git anchoring, optional stack manifest.

When to use

User asks for second brain, LLM wiki, vault ingest, Obsidian + Ollama, or cites Biophase7 lygo-second-brain.zip
Compounding notes from raw/ → permanent/ → wiki/ with retrieval (embed_index.py / search.py)
Consensus before saving contentious claims (consensus.py)

When not to use

Cloud-only Kimi/Claude as primary (optional fallback only)
Autonomous overnight /dream v1 (not shipped — build after ingest+index stable)
Auto git push, ClawHub publish, or kernel egg plant without consent

Setup

npx clawhub@latest install deepseekoracle/lygo-second-brain
export LYGO_STACK_ROOT=/absolute/path/to/lygo-protocol-stack
export LYGO_VAULT_ROOT=$LYGO_STACK_ROOT/lygo_second_brain/vault
python tools/install_lygo_second_brain.py
ollama pull llama3.2 && ollama pull nomic-embed-text

Commands (agents)

Intent	Command
Ingest	`python tools/lygo_second_brain.py ingest vault/raw/file.md`
Index	`python tools/lygo_second_brain.py index`
Search	`python tools/lygo_second_brain.py search "topic"`
Consensus	`python tools/lygo_second_brain.py consensus "question?"`
Wiki	`python tools/lygo_second_brain.py wiki "topic"`
Self-check	`python scripts/self_check.py` (from skill mirror)

Direct scripts: $LYGO_STACK_ROOT/lygo_second_brain/scripts/

LYGO lattice (honest)

Integrity: file read + size cap on ingest (not content entropy scoring)
Memory: git commits + optional data/second_brain/manifest.jsonl
Consensus: 2+ Ollama models, cosine similarity on answers
Optional: lygo-kernel-egg-planter for permaweb pins (user consent)

Skill chain

lygo-protocol-stack-operator → lygo-second-brain → lyra-brain / book-brain → lygo-kernel-egg-planter (optional)

Maintainer

npx clawhub@latest publish "…/clawhub/mirrors/lygo-second-brain" --slug lygo-second-brain --name "LYGO Second Brain"

Δ9Φ963 — local vault, verify, then compound.