Memory

BOOK BRAIN – LYGO 3-Brain Filesystem Helper

Build an additive filesystem library with daily indexes, topic files, state indexes, and outer reference stubs.

What it does

Scaffold and maintain a durable LYRA 3-Brain filesystem for OpenClaw/ClawHub Havens and Builder Key workspaces. It inspects the existing tree, previews an additive layout with dry-run, and creates approved folders, seed indexes, daily entries, and `.ref.txt` pointers. Topic details stay in library files while daily indexes and outer stubs remain short and free of secrets.

When to use it

  • Scaffolding a fresh OpenClaw or ClawHub Haven
  • Auditing a messy workspace without mass moves
  • Building indexes for long-horizon retrieval
  • Recording session, lattice, egg, or Builder Key references

The skill document

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.

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 retrieval — state/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-.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.txt → verify_lattice_alignment.py outputs
Kernel eggsreference/LYGO_KERNEL_EGGS.ref.txt — see lygo-kernel-egg-planter
Champion councilreference/LYGO_CHAMPION_COUNCIL.ref.txt → lygo-champion-council
ClawHub catalogreference/CLAWHUB_DEEPSEEKORACLE.ref.txt — 42 skills @ deepseekoracle
Builder Keyreference/BUILDER_KEY.ref.txt → ARCHITECT_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)

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-brain ↔ lyra-brain → lygo-protocol-stack-operator → lygo-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.

Questions people ask

Will it overwrite or reorganize my existing workspace?
No by default. It reports what exists or is missing, proposes changes with a dry-run, and only creates approved missing items; differing layouts can use a new branch or dated index.
What files and folders does it manage?
The recommended layout includes `memory/`, `reference/`, `brainwave/`, `state/`, `logs/`, `tools/`, and `tmp/`. It can seed `INDEX.txt` and `state/memory_index.json`, append daily indexes, and write `.ref.txt` outer pointers.
How does it work with lyra-brain?
BOOK BRAIN manages the on-disk library and outer reference stubs. At session close, it can optionally hand off a topic snip to lyra-brain's `session_log_snip.py` for graph growth.

Related skills

Brainstorming

Work through uncertain ideas iteratively to clarify options, tradeoffs, and a strategy.

129 installs7 stars

soul-in-sapphire

Preserve inspectable state, memory, journal, and identity continuity across sessions with Notion-backed records.

82 installs

Cloud Backup

Encrypt, upload, verify, retain, and restore scoped OpenClaw backups in S3-compatible storage.

89 installs4 stars

Journal

Capture entries, restart a lapsed practice, review past writing, and assess recurring themes.

71 installs2 stars

openclaw-nextcloud

Manage Nextcloud notes, tasks, calendars, files, contacts, shares, and Deck boards from one JSON CLI.

187 installs9 stars