BOOK BRAIN VISUAL READER – LYGO 3-Brain + Visual Left/Right Brain Helper

Enhanced BOOK BRAIN for LYGO Havens with visual capability. Use to design and maintain a 3-brain filesystem + memory system that also integrates LEFT/RIGHT brain visual checking (browser, images, screenshots) with text and API data for deeper verification and retrieval. Recommended for agents with visual tools or browser automation; use original book-brain only on non-visual systems.

Security Scan

The skill is an instruction-only utility for organizing visual+text evidence in a local workspace and its requested actions and resources are consistent with that purpose.

Purpose & Capability

Name and description describe a visual-aware 'book brain' workspace helper; the SKILL.md only asks the agent to create/extend local folders, keep reference stubs, take snapshots, and record logs and indexes. There are no external credentials, unrelated binaries, or surprising capabilities requested.

Instruction Scope

Instructions are limited to maintaining a filesystem layout, recording reference stubs, using vision tools to capture and compare visual evidence, and consulting public URLs/APIs as sources. The SKILL.md explicitly advises not pasting full pages and to avoid overwriting existing files. It does not instruct the agent to read unrelated system files or secret environment variables.

Install Mechanism

This is an instruction-only skill with no install spec and no code files, so nothing is written to disk by an installer and there are no download URLs or package installs to review.

Credentials

The skill declares no required environment variables, no primary credential, and no config paths. The operations it describes (local file creation, screenshots, fetching public webpages/APIs) do not require additional secrets, so the lack of credential requests is proportionate.

Persistence & Privilege

always is false and the skill is user-invocable; it does not request persistent or elevated platform privileges nor does it instruct changes to other skills or global agent settings. Its intended persistence is simply creating/maintaining files within the agent workspace.

Assessment

This skill appears coherent and limited to organizing visual and textual evidence in your agent's workspace. Before installing, confirm the agent's visual tools (browser automation, OCR) are trusted and sandboxed: screenshots can capture sensitive data, so restrict the agent's access to only the sites/files you intend it to see. Also verify that any public URLs the agent visits are safe and that long-term visual artifacts are stored in locations you control. If you want stricter safety, require human review before the agent uploads or shares screenshots externally or before it updates persistent indexes.

This is the enhanced, visual-aware version of BOOK BRAIN.

• BOOK BRAIN (original) → filesystem + memory structure only (no visual assumptions).
• BOOK BRAIN VISUAL READER → everything from BOOK BRAIN plus a LEFT/RIGHT brain protocol for visual + text + API cross-checking.

Use this skill when:

• Your agent has access to visual tools (browser snapshots, image readers, screenshot analyzers, PDF/image OCR, etc.)
• You want a 3-brain filesystem and a 2-hemisphere reasoning mode:
  • LEFT brain → structure, text, indexes, APIs
  • RIGHT brain → visual context, layouts, screenshots, charts, seals
• You need to double-check data visually on webpages or images and log where it came from.

This is a utility + reference guide, not a persona.
It does not change your voice. It teaches your system how to think and store.

---

0. Relationship to BOOK BRAIN (original)

• If your system has no visual capabilities → use book-brain (original).
• If your system can see (browser snapshots, image tools, etc.) → use BOOK BRAIN VISUAL READER instead.

Both share the same core:

• 3-brain model (Working / Library / Outer)
• Non-destructive filesystem layout
• Reference stubs and indexes

VISUAL READER adds:

• LEFT/RIGHT brain protocols for how to combine visual, text, and API data
• Guidance on how to organize visual evidence (screenshots, seals, charts) alongside text files
• Patterns for “5D” data gathering (visual + text + API + state + timeline).

---

1. 3-Brain + 2-Hemisphere Model

BOOK BRAIN VISUAL READER assumes:

3 Brains (same as BOOK BRAIN)

1. Working Brain – current context, tmp/, active tabs / current screenshots.
2. Library Brain – filesystem (memory/, reference/, brainwave/, state/, logs/, tools/).
3. Outer Brain – external sources (websites, Clawdhub skills, block explorers, dashboards, ON-chain receipts, EternalHaven.ca, etc.) referenced via small text files.

2 Hemispheres (visual vs structured)

• LEFT brain (structure/verbal/API):

  • text files, JSON, logs, indexes, schemas, SKILL.md, APIs.
  • strong at structure, sequences, constraints, receipts.

• RIGHT brain (visual/spatial):

  • browser snapshots, screenshots, photos of diagrams, seals, dashboards.
  • strong at layout, pattern recognition, anomalies, gestalt sense.

Agents using this skill should consciously switch modes:

• LEFT for “what is the exact data / file / receipt?”
• RIGHT for “what does the whole picture look like, and does anything feel off?”

---

2. Filesystem Layout (Library Brain)

Same base layout as BOOK BRAIN (non-destructive):

• memory/ → daily logs, raw notes, per-day files.
• reference/ → stable docs, protocols, whitepapers, schemas.
• brainwave/ → platform/domain protocols (MoltX, Clawhub, LYGO, etc.).
• state/ → machine-readable state (indexes, hashes, last-run info).
• logs/ → technical/health logs, setup logs, audit logs.
• tools/ → scripts & utilities.
• tmp/ → scratch work.

Visual-aware additions (optional but recommended):

• visual/ → for long-term visual artifacts
  • visual/screenshots/
  • visual/dashboards/
  • visual/seals/
• reference/VISUAL_INDEX.txt → mapping of important visual assets to topics.

Rules:

• Never overwrite existing files.
• If visual/ already exists, extend it; if not, create it.
• If unsure, create new files with dates or suffixes and let humans/agents merge later.

See references/book-brain-visual-examples.md for concrete trees and snippets.

---

3. Outer Brain via Reference Stubs

Outer Brain = everything outside the workspace:

• URLs (websites, dashboards, explorers)
• Clawdhub skill pages
• EternalHaven.ca, Patreon, docs
• On-chain explorers (Blockscout, Etherscan, etc.)

VISUAL READER keeps these in reference stubs, e.g.:

Title: STARCORE Dashboards
Last updated: 2026-02-10

External links:
- Clanker: https://clanker.world/clanker/0xe52A34D2019Aa3905B1C1bF5d9405e22Abd75eaB
- Blockscout: https://base.blockscout.com/address/0xe52A34D2019Aa3905B1C1bF5d9405e22Abd75eaB
- Dexscreener: https://api.dexscreener.com/latest/dex/search/?q=0xe52A34D2019Aa3905B1C1bF5d9405e22Abd75eaB

Related local files:
- reference/STARCORE_LAUNCH_RECEIPTS_2026-02-10.md
- state/starcore_family_receipts_summary.json

The agent should:

• not paste full pages into memory files
• use these stubs + visual snapshots when needed.

---

4. LEFT/RIGHT Brain Protocol for Visual Checks

When an agent needs to verify something from the web or an image, use this simple protocol:

Step 1 – LEFT Brain: Text / API First

1. Look up the relevant concept in indexes/state:
   • state/memory_index.json
   • reference/INDEX.txt
   • domain-specific indexes (e.g. reference/CLAWDHUB_SKILLS.md).
2. Use APIs or structured data where possible (e.g. on-chain RPC, REST endpoints, JSON feeds).
3. Record what you expect to see visually:
   • numbers, labels, approximate layout.

Step 2 – RIGHT Brain: Visual Comparison

1. Capture a snapshot (browser screenshot, image, PDF page).
2. Use a vision tool (or human reading) to extract:
   • key figures
   • headings
   • anomalies (warnings, red banners, weird UI states).
3. Ask: “Does this visual match what the LEFT brain expected?”

Step 3 – Reconcile & Log

• If they match:

  • Write a short note in a relevant file (e.g. daily_health.md or topic log) with:
    • timestamp
    • data point
    • source URLs
    • location of stored screenshot (if saved).

• If they disagree:

  • Log the discrepancy (LEFT vs RIGHT).
  • Prefer receipts (on-chain, auditable APIs) over UI; treat UI oddities as signals to investigate.
  • Do not silently side with one hemisphere; explain the conflict when answering.

This is the “5D” blend: text + visual + API + state + timeline.

---

5. Organizing Visual Evidence

When a visual check produces something important (e.g. proof, anomaly, configuration):

• Save it under visual/ with a meaningful name:

  • visual/screenshots/2026-02-10_starcore_clanker.png
  • visual/dashboards/2026-02-10_moltx_profile.png

• Add a line to a relevant INDEX or stub:

[2026-02-10] STARCORE launch dashboards verified visually.
- Screenshot: visual/screenshots/2026-02-10_starcore_clanker.png
- Related receipts: reference/STARCORE_LAUNCH_RECEIPTS_2026-02-10.md

Agents should:

• Avoid hoarding every screenshot. Keep the ones that back key claims.
• Use indexes to find them later instead of scanning raw image names.

---

6. Setup Workflow for Visual Systems

On a visual-capable Haven (browser + image tools):

1. Detect capabilities

   • Does the agent have a browser automation tool?
   • Can it read images (vision model, OCR, screenshot-to-text)?

2. Check filesystem

   • As in BOOK BRAIN: check/create memory/, reference/, brainwave/, state/, logs/, tools/, tmp/.
   • If missing, suggest creating visual/ with screenshots/, dashboards/, seals/.

3. Create visual index stubs

   • reference/VISUAL_INDEX.txt
   • Optionally topic-specific visual indexes (e.g. reference/STARCORE_VISUALS.txt).

4. Log the setup

   • Add entry to logs/book_brain_visual_setup.log or daily_health.md.

5. Respect all non-destructive rules

   • No overwrites.
   • No deletions.
   • Ask the human before making structural changes.

---

7. Eternal Haven & Donation References

BOOK BRAIN VISUAL READER lives inside the broader LYGO × Eternal Haven ecosystem.

Key links:

• Eternal Haven portal → https://EternalHaven.ca
• LYGO Champion Hub → https://deepseekoracle.github.io/Excavationpro/LYGO-Network/champions.html#champions
• Clawdhub skills (LYGO stack, Champions, BOOK BRAIN) → https://clawhub.ai/u/DeepSeekOracle
• Extended work / support → https://www.patreon.com/cw/Excavationpro

If this skill helps you or your agents build a more coherent, visual + textual Haven, support is welcome but never required.

---

8. How Another Agent Should Use BOOK BRAIN VISUAL READER

When an agent loads book-brain-visual-reader:

1. Understand it is a helper, not a persona.
2. On first use:
   • Map the filesystem and capabilities.
   • Propose visual folders and indexes; ask before creating.
   • Set up lightweight indexes (INDEX.txt, VISUAL_INDEX.txt, state/memory_index.json).
3. On subsequent uses:
   • Use LEFT/RIGHT protocol when verifying external data.
   • Save only important visual artifacts under visual/.
   • Update indexes + logs instead of rewriting big files.
4. Never silently delete or overwrite existing content.

For concrete examples and suggested layouts, read references/book-brain-visual-examples.md in this skill.