docs/AI-COGNITIVE-LOGGING-PROTOCOL.md

🎯 1. The Purpose of the Immutable Ledger

In the DSOM framework, the Spatial Memory Palace (.agents/brain/) contains distilled, human-readable summaries written by the AI and the Architect during the EOD Hibernation. However, human-readable summaries lack bit-perfect forensic detail (e.g., exact bash commands executed, regex applied, raw error outputs). To solve this, we integrate the AI’s native Conversation Transcripts (System Logs) into the Palace as an Immutable Ledger.

🌐 2. Cross-AI Compatibility Mandate

Because DSOM is provider-agnostic, you may be Gemini, Claude, OpenAI, or a local open-source LLM. Every AI provider has a different way of storing conversation history. If you are not Gemini:

You MUST document how your specific provider logs conversation history (e.g., where your transcripts are stored, or if you only use a cloud dashboard).
You MUST explain your logging mechanism to the User and to other AIs (like Gemini) by creating a section in your relevant closet.md or by updating this document to include your specific format.
You MUST establish a way to map your specific session/conversation ID to the Palace.

🔗 3. How to Map Logs to the Palace

When updating a closet.md during the EOD Hibernation Ritual, the active AI must inject its specific Session/Conversation ID and Log Path at the top of the file.

Standard Format Header

# 🗄️ Spatial Closet: [Topic]

**Primary Conversation ID:** [Insert Unique Session/Conversation ID Here]
**Log Path / Reference:** [Insert Absolute Path to Transcript OR Cloud URL]

## 📌 Context

[The distilled human-readable summary...]

🧠 4. Current Implementations

Implementation A: Google DeepMind Gemini (Antigravity)

Mechanism: Gemini automatically generates chronological JSONL transcripts for every conversation.
Log Location: <appDataDir>\brain\<conversation-id>\.system_generated\logs\transcript.jsonl
Data Stored: Exact user prompts (USER_INPUT), terminal commands executed (tool_calls), and thought processes.
How it maps: We drop the Gemini conversation-id into the closet.md. If a future agent needs to know exactly what bash command Gemini ran to fix a bug, they grep that specific transcript.jsonl file.

Implementation B: [Other AI Provider]

(If you are Claude, ChatGPT, or another agent reading this: You must fill out this section for your own architecture when you encounter it!)

Maintained by the DSOM Engineering Team | Cross-AI Logging Standard v1.0

Deep State of Mind (DSOM) For My AI Protocol | Harisfazillah Jamel (LinuxMalaysia) | 2026-07-04 Standard: UK English | DBP-standard Bahasa Melayu Malaysia (Piawai) | GNU General Public License v3.0

Reference

Explanation

Governance

📜 AI Cognitive Logging Protocol (Cross-AI Standard)

docs/AI-COGNITIVE-LOGGING-PROTOCOL.md

🎯 1. The Purpose of the Immutable Ledger

🌐 2. Cross-AI Compatibility Mandate

🔗 3. How to Map Logs to the Palace

Standard Format Header

🧠 4. Current Implementations

Implementation A: Google DeepMind Gemini (Antigravity)

Implementation B: [Other AI Provider]

​docs/AI-COGNITIVE-LOGGING-PROTOCOL.md

​🎯 1. The Purpose of the Immutable Ledger

​🌐 2. Cross-AI Compatibility Mandate

​🔗 3. How to Map Logs to the Palace

​Standard Format Header

​🧠 4. Current Implementations

​Implementation A: Google DeepMind Gemini (Antigravity)

​Implementation B: [Other AI Provider]

docs/AI-COGNITIVE-LOGGING-PROTOCOL.md

🎯 1. The Purpose of the Immutable Ledger

🌐 2. Cross-AI Compatibility Mandate

🔗 3. How to Map Logs to the Palace

Standard Format Header

🧠 4. Current Implementations

Implementation A: Google DeepMind Gemini (Antigravity)

Implementation B: [Other AI Provider]