Why publish a local control plane?
RamLabs AI Dashboard is the local operational view for Codex, Claude Code, Hermes, and OpenClaw. It combines fleet status, terminal links, analytics, backup readiness, capacity telemetry, and a metadata-only Activity ledger. Publishing its source gives the dashboard a reviewable recovery path without treating a deployed VM directory as a backup artifact.
What made the repository safe to publish?
- Canonical application source, static assets, tests, and configuration schema were included.
- Runtime state was excluded: audit events, alert state, analytics snapshots, backup status, service logs, and rollback copies.
- Python bytecode and obsolete frontend backup files were removed and ignored.
- A repository README documents local validation, deployment ownership, and LAN-only security boundaries.
Operational contract
The dashboard presents backup freshness and restore readiness separately. A completed copy or GitHub push is not a restore test. The Activity ledger records bounded action metadata and digests—not submitted prompt text, secrets, file contents, or backup plaintext.
Validation before publication
python3 -m py_compile app.py dashboard_ops.py node --check static/app.js pytest -q git diff --cached --check
What is intentionally not in Git?
Secrets, environment files, runtime audit data, raw operational logs, backup payloads, and local recovery material remain outside version control. This keeps the repository useful for code recovery while avoiding an accidental export of the VM control plane.