Skip to content

🀝 HUMAN-HANDOVER-CONTEXT.md

Path: docs/HUMAN-HANDOVER-CONTEXT.md

DSOM For My AI β€” SESSION HANDOVER TEMPLATE Adapt this file for every new project cloned from the DSOM skeleton. Fill in every [PLACEHOLDER] with your project-specific values. This file is the first document you paste into a new AI session.


⚑ SESSION HANDOVER PROMPT

(Copy everything below this line and paste as your first AI message)


I am starting a new AI session. Synchronise your state with the following:

  1. READ the project brief: docs/ directory and the DSOM brain files.
  2. SYNC with the task tracker: .agents/brain/task.md
  3. READ brain artifacts: .agents/brain/walkthrough.md and .agents/brain/implementation_plan.md
  4. VERIFY docs/AI-COGNITIVE-TWIN-PROTOCOL.md β€” confirm all [PLACEHOLDER] fields are filled.
  5. READ README.md, SUMMARY.md, CHANGELOG.md
  6. RUN git log --oneline -20 to read recent commit history.
  7. WALK THE PALACE β€” read .agents/brain/palace_registry.md (Section [14] of the SOD manifest) and identify the relevant Wings and Rooms for today's work.

Run Start of Day (SOD) as per DSOM For My AI Protocol v6.1 + Palace v1.0.

T2 shortcut: bash tools/sod-palace.sh # (Windows: .\tools\sod-palace.ps1) automates steps 1–6 + palace check automatically. Upload the manifest and use the handshake phrase below.


πŸ—οΈ Project Identity

  • Project Name: [YOUR_PROJECT_NAME]
  • Stack: [YOUR_STACK e.g. Logstash + Kafka + Podman / ELK / Ansible roles]
  • Purpose: [ONE_LINE_PURPOSE]
  • Repo: [YOUR_GITHUB_REPO_URL]
  • Protocol Version: DSOM v6.1 + Palace v1.0 | GitOps Β· AIOps Β· Ansible

πŸ—ΊοΈ 4-Tier Command Highway

We strictly follow the 4-Tier Command Highway (T1β†’T2β†’T3β†’T4):

Tier Name Environment Path
T1 Command Centre Windows 11 + PowerShell + Google Antigravity [T1_WINDOWS_PATH]
T2 Dev Bridge WSL2: dsom-control-almalinux10 /mnt/[T2_MOUNTED_PATH]
T3 Jump Host / Ansible Control [T3_HOST_IP_OR_NAME] /opt/[YOUR_PROJECT_NAME]/
T4 Production Fabric [T4_NODE_IPS_OR_NAMES] [T4_TARGET_PATH]

Code is promoted: T1 β†’ Git β†’ T2 β†’ T3 β†’ T4. No shortcuts.


πŸ›‘οΈ Security Doctrine: "Rootful Control, Rootless Application"

  • Rootful Control: Ansible/sudo manages OS-level tasks β€” systemd, networking, filesystem.
  • Rootless Application: All containers/services run as [APP_USER] (UID [APP_UID] GID [APP_GID]).
  • Sovereign Identity: dsom-admin:2001:2001 on target nodes.
  • Dev Identity: [DEV_USERNAME]:[DEV_UID]:[DEV_GID] on WSL2.

βš™οΈ Environment Rules

  • AI Partner: Google Antigravity on Windows 11 (PowerShell).
  • Ansible: Runs from T2 WSL2 (dsom-control-almalinux10). NOT from Windows directly.
  • Git operations: All git add, git commit, git push done on T1 Windows/PowerShell.
  • Dev testing: Human runs scripts inside T2 WSL2 (dsom-control-almalinux10). AI guides and waits for output.
  • Production:hosts.prod.yml is in .gitignore. Human must supply it for any production Ansible run.

πŸ”„ The Safe Sync Ritual (Run After Every T1 Push)

Whenever changes are pushed from Windows (T1), perform this ritual to align all tiers:

Step 1 β€” Pull on T2 (WSL2: dsom-control-almalinux10)

cd /mnt/[T2_MOUNTED_PATH]
git pull origin main

Step 2 β€” Promote to T3 Jump Host / Ansible Control (if applicable)

# Run from T2 WSL2 β€” promote project to T3 orchestrator

rsync -avz --checksum --delete \
    --exclude '.git' --exclude '.gemini' --exclude '.agents' \
    ./ dsom-admin@[T3_IP]:/opt/[YOUR_PROJECT_NAME]/

Step 3 β€” Verify T3 Ansible Connectivity

# Run from T2 WSL2

ansible all -m ping -i inventory/hosts.yml

Step 4 β€” Optional: Remote Audit from T1 via WSL

# Run from Windows PowerShell (T1) β€” bridges through T2 β†’ T3

wsl -d dsom-control-almalinux10 -u dsom-admin -e bash -c `
    "ssh -t dsom-admin@[T3_IP] 'cd /opt/[YOUR_PROJECT_NAME] && bash tools/audit-pre-flight.sh'"

🏁 Sovereign Directive (The Handshake Trigger)

After the AI reads all context, trigger the session with:

"Initialise DSOM Protocol v6.1 + Palace v1.0 for [YOUR_PROJECT_NAME]. Read the uploaded manifest or this handover file. Walk the Palace Registry in Section [14] β€” identify the relevant Rooms for today's work. Confirm the 4-Tier Command Highway is mapped. Verify docs/AI-COGNITIVE-TWIN-PROTOCOL.md is filled in. State the last Mental Anchor from .agents/brain/walkthrough.md. Respond in UK English. State: 'Sovereign State Synchronised β€” [YOUR_PROJECT_NAME] is live.' when ready."


πŸ“‹ How to Provide Output for AI Review

Method Command
Direct Paste Copy full terminal output into AI chat
Log File bash tools/audit-pre-flight.sh \| tee audit_result.log β€” share log content
Ansible Output ansible-playbook playbooks/dsom/audit-preflight.yml -i localhost, \| tee audit_ansible.log

AI will wait for your results before proceeding. Never assume success.


πŸ“Œ Current State (Fill in at each session end)

CURRENT STATE: [DESCRIBE_CURRENT_STATE e.g. v1.0 β€” Ansible baseline deployed, Kafka brokers up]
Environment:   [e.g. 3-node test cluster]
Last Commit:   [PASTE LAST COMMIT HASH AND MESSAGE]
Mental Anchor: [ONE SENTENCE β€” exact state and where to resume]
Next Action:   [WHAT THE AI SHOULD DO FIRST IN THE NEW SESSION]

πŸ”‘ Key Reminders to AI Agent

  1. Advisory Mode only. Propose β†’ Human approves β†’ Ansible executes β†’ AI verifies.
  2. Never run commands on production nodes. Write Ansible playbooks. Human runs them.
  3. Every step is committed to Git before execution. No Git = no change.
  4. Idempotency is mandatory. Every playbook must be safe to re-run.
  5. All secrets via ansible-vault. Never commit plaintext secrets.
  6. Detailed Git commit messages. Format: type(scope): description [Phase/vXXX]
  7. UK English only in all documentation and responses.
  8. Wait for human output before proceeding with the next step.
  9. Control yourself during long debugging. Emit one action at a time.
  10. Both T1 and T2 must be in sync β€” always pull on T2 after T1 push.
  11. Walk the Palace Registry at the start of every session β€” read palace_registry.md and load the relevant closet.md files before advising.
  12. Propose Palace closet updates at EOD if new decisions or discoveries were made.

Standard: DSOM For My AI Protocol v6.1 + Palace v1.0 | Harisfazillah Jamel | LinuxMalaysia Template source: docs/HUMAN-HANDOVER-CONTEXT.md | Updated: 2026-04-08 Adapted from: DSOM elasticsearch-kibana-podman project (v9.4 Sovereign Recovery model)


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