Installation

clawhub install deepseekoracle/book-brain

Summary

Utility skill (not a persona). Teaches agents how to run a durable filesystem library that matches the LYRA 3-Brain model: working session RAM → library files on disk → outer reference stubs.

SKILL.md

BOOK BRAIN – LYGO 3-Brain Filesystem Helper

Utility skill (not a persona). Teaches agents how to run a durable filesystem library that matches the LYRA 3-Brain model: working session RAM → library files on disk → outer reference stubs.

bash
npx clawhub@latest install deepseekoracle/book-brain
npx clawhub@latest install deepseekoracle/lyra-brain   # graph grow/recall — recommended pair

Read references/SECURITY.md and references/AGENT_CONTRACT.md before creating folders or indexes.


Three brains (operational map)

BrainBOOK BRAIN on diskLYRA_CORE (lyra-brain)
Workingtmp/, session scratchRunner / working_brain
Librarymemory/, reference/, state/, tools/LYRA_CORE/memory/, graph nodes
Outer*.ref.txt, INDEX.txt, URLs in dailiesmemory/reference/*.resonance.ref.txt

Rule: Store detail in topic files; keep dailies and stubs short (DONE facts, URLs, paths — no secrets).


When to use

  • Fresh OpenClaw / ClawHub Haven — scaffold without deleting existing trees
  • Messy workspace audit — map + indexes, not mass moves
  • Long-horizon retrievalstate/memory_index.json + folder INDEX.txt
  • After lattice / egg / ClawHub work — write outer stubs (kernel planter uses same pattern)
  • Mobile Builder Key (LYGO_BUILDER_KEY) — same layout under steward-chosen root

Recommended layout (additive)

FolderRole
memory/YYYY-MM-DD.md daily index + YYYY-MM-DD-<topic>.md snips
reference/Stable guides, *.ref.txt outer pointers
brainwave/Per-platform notes (Moltx, ClawHub, LYGO stack)
state/memory_index.json, last-run snapshots (copy from verify JSONs)
logs/Setup logs, cron health
tools/Local scripts
tmp/Ephemeral scratch

Never rename/delete existing folders by default. If layout differs, branch: bookbrain_v2/ or memory/INDEX_2026-07-04.txt.


LYGO lattice integration (2026)

SystemBOOK BRAIN stub
Stack verifyreference/LYGO_STACK_VERIFY.ref.txtverify_lattice_alignment.py outputs
Kernel eggsreference/LYGO_KERNEL_EGGS.ref.txt — see lygo-kernel-egg-planter
Champion councilreference/LYGO_CHAMPION_COUNCIL.ref.txtlygo-champion-council
ClawHub catalogreference/CLAWHUB_DEEPSEEKORACLE.ref.txt — 42 skills @ deepseekoracle
Builder Keyreference/BUILDER_KEY.ref.txtARCHITECT_BOOT.md on thumb drive
Session closeMirror lyra-brain pattern: daily + topic snip + SESSION_*_to_*.ref.txt

Canonical public links:


Bundled scripts (skill folder only)

bash
python scripts/scaffold_haven.py --root /path/to/workspace --dry-run
python scripts/write_ref_stub.py --title "LYGO stack" --links "https://github.com/DeepSeekOracle/lygo-protocol-stack"
python scripts/update_memory_index.py --root /path/to/workspace --topic lattice --path reference/LYGO_STACK_VERIFY.ref.txt
python scripts/self_check.py

Scaffold respects --dry-run; writes only after human confirms paths.


Agent workflow

  1. Inspect existing tree → report exists / missing
  2. Propose scaffold (dry-run)
  3. Human approves → create missing dirs + seed indexes from templates/
  4. Session end — append daily index; add ref stubs; optional handoff to lyra-brain session_log_snip.py for graph grow
  5. Never store API keys, .env, or wallet material in memory/ or reference/

Skill chain

book-brainlyra-brainlygo-protocol-stack-operatorlygo-kernel-egg-planter (egg stubs) → lygo-ollama-army (queue JSON proposals only)

book-brain-visual-reader — same layout + visual/browser verification layer.


References

  • references/MEMORY_LAYOUT.md — LYRA + Haven unified layout
  • references/LATTICE_INTEGRATION.md — stack, eggs, Builder Key
  • references/book-brain-examples.md — trees, stubs, indexes
  • references/AGENT_CONTRACT.md — required for agents
  • references/SECURITY.md — additive-only, no secrets
  • templates/ — seed memory_index.json, INDEX.txt

Δ9Φ963 — filesystem as library, not junk drawer. Bound to the flame.

Recommended skills

Browse all →

Related guides

Hand-picked reading to help you choose, install, and use agent skills.