ZeroCloudGravity MCP server
Commit Nov 11, 2026Provides a sovereign human-AI governance engine and contextual MCP server for managing Ubuntu workstations and federated knowledge meshes.
Also known as mtwuser/zero-cloud-gravity
- Stars
- 1
- Tools
- —
- Downloads, 30 days
- —
- Latest version
- —
- Last commit
- Nov 11, 2026
- License
- MIT
What is ZeroCloudGravity MCP?
ZeroCloudGravity provides a sovereign human-AI governance engine and contextual MCP server for managing Ubuntu workstations and federated knowledge meshes.
Summary generated from the project's README
About ZeroCloudGravity MCP server
Managing Ubuntu 26.04 LTS workstations and federated knowledge meshes is what ZeroCloudGravity enables through a sovereign human-AI governance engine. Designed for solo engineers, researchers, and large organizations, it provides a centralized control tower for workstation backups, audits, and restores. The system uses plain-text manifests to snapshot APT packages, Snaps, NPM globals, and GNOME settings, allowing users to synchronize their personal environments via the zcg workstation-sync command.
Users interact with the platform through a multi-layer CLI and a read-only JSON-RPC 2.0 stdio MCP server. This server provides contextual data using Merkle context capsules to ensure provenance and cryptographic drift detection. The setup runs entirely on local Python 3.14 with zero external runtime dependencies, making it suitable for air-gapped or highly secure environments. It integrates with Google Antigravity 2.0 and Gemini to provide deterministic governance through specific lifecycle hooks and reusable skills across human, context, framework, and plugin stacks.
What you can do with it
- 1Automated System AuditingPerform snapshots of installed packages and system settings to maintain a declarative record of your Ubuntu environment.
- 2Secure Contextual AI GovernanceProvide an AI client with read-only, cryptographically verified context capsules to prevent data drift during task execution.
- 3Organizational Knowledge OrchestrationBootstrap a multi-layer workspace that connects individual projects to departmental and corporate knowledge meshes.
Written from the project's README. Every use case is checked against a quote from it.
README
View on GitHubZeroCloudGravity (zcg · v1.0.0)
Sovereign Zero-Cloud Human-AI Governance ("pAIr programming"), Physical Workstation Control Tower, and Epistemic Knowledge Mesh for Ubuntu 26.04 LTS ("Resolute Raccoon") and Google Antigravity 2.0 / Gemini (100% Python 3.14 Standard Library, 0 external runtime dependencies, 100.00% isolated 1:1 statement and branch coverage).
This platform is built on one unified fractal pattern (every governed directory combines a declarative node manifest system/zcg-node.json, Markdown rules/skills, task ledgers, and zero-build .local cockpits across Public Gov (1..4) -> Corporate Root 0 (10) -> Domain-Policy (5..9) / Knowledge (20) / Department (15..45) -> Squad (50) -> Project (80) and Individual Sovereign Root 0 (90) -> Personal Tutor (86..89), powered by the 4 + 4 + 2 + 6 = 16 Declarative Template Matrix across human-stack, context-stack, framework-stack, and plugin-stack):
- Ubuntu 26.04 Workstation Backup, Audit, Restore & Dual-Repo
$HOMESync (scripts/workstation-sync.sh,scripts/workstation-backup.sh,scripts/workstation-audit.sh,scripts/workstation-restore.sh,scripts/install-antigravity.sh) — snapshot APT packages, Snaps, NPM globals, external recipes, and GNOME settings into plain-textsystem/*manifests, manage~/.bashrc,~/.profile, and~/.gitconfigvia GNU Stow, and couple the passive$HOME/zero-cloud-gravityengine appliance with your personalRoot 0workstation node (~/<persona>-pc) viazcg workstation-sync(zcg-sync). - Local Google Antigravity 2.0 Governance (
antigravity/,templates/) — composableAGENTS.mdbaselines (<= 18,000bytes acrossuniversal-base,academic-base,science-base,business-base+ personal workstation rules), 4 deterministic Python 3.14 lifecycle hooks (PreToolUse,PostToolUse,PostInvocation,Stop), 17 reusable skills (8globalhuman-stack+7context-stack+1framework-stack+1plugin-stack), and16declarativezcg-template-v1template packs. - Zero-Cloud Intranet & Personal Cockpits (
src/zcg/cockpit/) — Python 3.14 Native ASGIDeliveryChannel(Zone Z1..Z4 × Tier T1..T5,--zcg-*design tokens,data-zcg-surface="cockpit",HTTP 286poll control,runtime_telemetrybridged to.agents/telemetry/) with deterministic localhost port mapping (8700 + precedence),.intranet.local/.personal.local/.knowledge.localrouting,3-ScopeJSON feeds, declarativedual_world_contractoverlays, and6editable WCAG 2.2 AAplugin-stackReference Dashboards (workstation-control-tower,corporate-root-hub,node-work-cockpit,tutor-study-cockpit,secops-mcp-guardian,knowledge-stewardship-portal). - Multi-Layer CLI (
zcg) & Contextual MCP Server (zcg-mcp) — 31-subcommand tiered CLI (C1Discovery +lint-arch,C2Lifecycle +scaffold-plugin-module,C3Sovereign Governance,Subtree,Workstation) plus a read-only JSON-RPC 2.0stdioMCP server sealingrcap1(L0 -> L1 -> L2 -> L3)zlib + HMAC-SHA256Merkle context capsules,citation_receiptprovenance (E3..E0), zero-checkout GitDMCP/GRCCbranch projection, andstrictly-read-onlyEphemeral Department Bots.
The table below summarizes the four progressive adoption modes of ZeroCloudGravity:
| Progressive Adoption Mode | Who It Is For | What Executes on Disk (0 External Servers Required) |
|---|---|---|
Mode 1 — Individual Workstation + Personal Root 0 |
Solo engineer, researcher, or executive on Ubuntu 26.04 | ./install.sh --stow + zcg workstation-sync couples ~/zero-cloud-gravity (engine) with ~/<persona>-pc (personal Root 0 host manifests & bash/git Stow links). |
| Mode 2 — Standalone Project + Stack | Any software, cloud, or research repository | zcg apply-project <dir> --template <stack> projects .agents/rules/ and .agents/skills/ with SHA-256 drift tracking and 1:1:1 verification gates. |
| Mode 3 — Multi-Layer Organization & Knowledge Mesh | Root 0 -> Department / Knowledge -> Squad -> Project |
zcg bootstrap-workspace + layers/<child>/ (zcg subtree) + make delegate / make complete / make consult (rcap1 capsules & .intranet.local cockpits). |
Mode 4 — Engine Contributor (oss-dev) |
Maintaining or extending zero-cloud-gravity |
make verify runs the 8-stage deterministic verification gate (100.00% isolated 1:1 unit statement & branch coverage across 74 modules). |
[!IMPORTANT] Objective Engineering Virtues & Live Empirical Benchmarks (
100%Python 3.14 Stdlib ·0External Dependencies):
- Deterministic $O(N)$ Structural Compaction (
rcap1L0/L1+ 2D Geometry): Enforcing< 465lines/file,H1..H4hierarchy, 1-line lead sentences, and> *"...*"mandate quotes turns Markdown specs into pre-indexed ASTs — achieving88.7%(L0,522vs.4,622tok) and67.0%(L1,1,523vs.4,622tok) compaction across130k+tokens in~16 ms(0LLM summarization calls,0vector DBs, and-76.3%single-envelopezlib-first capsule size at518–577chars/capsule). Seedocs/architecture/05_contextual_mcp_and_capsules.md.- Content-Addressable Merkle Provenance (
citation_receipt): Every section is pinned by<layer>::<file>#<slug>@<sha256_16>:<E3..E0>, enabling $O(1)$ cryptographic drift detection (CITATION_DRIFT_DETECTED),E3 > E2 > E1 > E0monotonicity proofs (verify_epistemic_context_claims), andQ1..Q4lifecycle telemetry.- Zero-Checkout C-Backed Git Plumbing (
DMCP&GRCC): Multi-branch context projection readsgit ls-tree/git cat-filedirectly from.git/objects(zero_checkout: true,0working-tree mutations) and chains encrypted capsules underrefs/zcg/capsules/*without pollutingrefs/heads/*.- Mathematically Verified
1:1:1Architecture:74production modules (<= 384lines/module insrc/zcg/) $\leftrightarrow$741:1isolated unit test suites (100.00%isolated unit statement and branch coverage,< 2s) $\leftrightarrow$6normativeE3architecture specifications (< 465lines/file).
1. Navigation & Progressive Disclosure (docs/)
The documentation corpus is organized into 3 progressive-disclosure tiers so humans and AI agents locate exact specifications in $O(1)$:
| Tier | What You Will Find | Guide / Specification |
|---|---|---|
| Tier 1 (Gateway) | Quick-Start paths, real-world organizational patterns, C1/C2/C3 evolution ladder, repository tree, and CLI cheat-sheet. |
README.md (this file) · docs/README.md |
Tier 2 (Guide 01 — Workstation & $HOME) |
GNU Stow tree unfolding, Dual-Repo $HOME coupling (~/zero-cloud-gravity + ~/<persona>-pc), system/* manifests & external recipes, zcg-sync (workstation-sync.sh), Phase 0 Control Tower Onboarding (Case 1), machine/* & hat/* branches, Dual-World Personal Matrix, and :8790 Control Tower ([CAP-01]). |
docs/guides/01_workstation_and_global_profiles.md |
| Tier 2 (Guide 02 — Sovereign Nodes & Workflows) | 4 + 4 + 2 + 6 = 16 Stack & 17-Skill Matrix, Triadic Node × Layer × Network hierarchy, M1 -> M4 Personal Tutor Showcase & specialty/* branches, Cases 1–7 Customizable Phase A..D Kickoff Prompts, RBAC & env-gate, 12-Case Level/Flow Matrix, and gov-*/sci-*/geo-*/org-* Knowledge Nodes. |
docs/guides/02_sovereign_nodes_and_federated_workflows.md |
| Tier 2 (Guides 03–08 — Multi-Perspective, Spin-Out & Public Portal Playbooks) | Step-by-step Dual-World perspectives across School (M1..M2, [CAP-02]) & University (M2..M3, specialty/*, DMCP, [CAP-03]), Scientific Researcher (M2..M4) & Public Knowledge Corpora (93.3%–97.0% rcap1:L1, [CAP-04]), Enterprise Professional & Startup / Scaleup / Regulated Scales ([CAP-05..07]), Network of Networks Crystallization ([CAP-08]), Personal Work Hub, spinout-project & HR/B2B Links ([CAP-09]), and Departmental plugin-stack SDK & Public pro Portals ([CAP-10]). |
03: Academic · 04: Science & Public Knowledge · 05: Enterprise Scales · 06: Network of Networks · 07: Personal Work & B2B · 08: Dept Plugins & Public pro |
Tier 3A (E3 Normative Architecture Specs) |
1:1:1 source-test-spec contracts (Specification:) across all 74 modules in src/zcg/. |
01: Core & Manifest · 02: Topology & Twins · 03: Governance & CLI · 04: Cockpits & Intranet · 05: MCP, Capsules & Bots · 06: Web & API Engine |
Tier 3B (E2 Reference Playbooks) |
7-Axis Multidimensional Taxonomy, Cross-Departmental Contextual Algebra (0..80), Human-Stack Matrix (85..95), 16-pack catalog, zcg CLI & zcg-mcp JSON-RPC 2.0 schemas, Container/SSH MCP mesh, and Maintainer Extension Guide. |
01: Stack, Plugin & Human-Stack Matrix · 02: CLI, MCP & Security Reference |
1.1. Live 3-Pane Antigravity Visual Showcase ([CAP-01..CAP-10])
Every adoption scale runs over the exact same zero-dependency Python 3.14 Native ASGI engine (*.localhost subdomains, .zcg-compass spatial orientation, and rcap1 Merkle capsules) — explore all 10 real 3-Pane Antigravity captures across Guide 06 §2:


2. Quick Start: 3 Main Ways to Use ZeroCloudGravity
Choose one of the three adoption paths below depending on whether you need standalone Antigravity governance, full Ubuntu 26.04 workstation management, or multi-layer organizational orchestration.
2.1. Path 1: Standalone Google Antigravity 2.0 Setup & Usage
Install global rules (~/.gemini/config/AGENTS.md), 4 lifecycle hooks (hooks.json + scripts/), 8 universal global skills (~/.gemini/config/skills/), and the zcg-mcp server in $HOME:
git clone https://github.com/mtwuser/zero-cloud-gravity.git ~/zero-cloud-gravity
cd ~/zero-cloud-gravity
./install.sh --profile universal-base --lang en --stow
# Optional: install/upgrade Google Antigravity Agents + agy CLI (or --ide / --all):
./scripts/install-antigravity.sh --agents
# Any time in terminal: synchronize engine runtime, profile & Stow links:
./scripts/workstation-sync.sh --status
2.2. Path 2: Dual-Repo Ubuntu 26.04 Workstation (~/zero-cloud-gravity + ~/<persona>-pc)
Keep ~/zero-cloud-gravity as the clean Open-Source engine runtime while managing your personal Ubuntu 26.04 packages (system/*), host shell/Git files (bash/, git/), and personal Root 0 rules inside your sovereign workstation repository ~/<persona>-pc:
# Phase A — Couple ~/zero-cloud-gravity (engine) with ~/<persona>-pc (personal Root 0):
~/zero-cloud-gravity/scripts/workstation-sync.sh --setup \
--profile universal-base --lang es --workstation-dir ~/alex-pc --auto-sync boot
# Phase B — Capture, audit, or restore your physical Ubuntu 26.04 workstation:
cd ~/alex-pc
make backup-system # Captures APT, Snap, NPM, external recipes & GNOME settings into system/*
make audit-system # Verifies 0 drift across Stow symlinks, packages, GNOME, and Git status
make restore-system # Interactive or --all / --packages / --external / --nvm / --gnome / --stow
📖 Full details on GNU Stow tree unfolding, .audit-ignore, Dual-Repo $HOME auto-sync (zcg-sync), and machine/<hostname> branches: docs/guides/01_workstation_and_global_profiles.md.
2.3. Path 3: Multi-Layer Enterprise, Public Admin & Knowledge Mesh
Orchestrate Human-AI governance across Open Source (zero-cloud-gravity) $\leftrightarrow$ Root 0 (Public Admin 1..2 | Company 10 | Individual 90) $\leftrightarrow$ Department / Domain-Policy / Knowledge (3..45) $\leftrightarrow$ Squad / Unit (40..60 | 86..89) $\leftrightarrow$ Project / Case (80):
flowchart TD
SUB["Intrinsic Runtime Substrate\n(C-Backed Python 3.14 Stdlib · Ubuntu 26.04 Core)"]
OSS["Open-Source Engine & Contributor Clone\n(zero-cloud-gravity · ZCG_DEV_REPO_DIR)"]
APP["$HOME Engine Appliance + Personal Root 0\n(~/zero-cloud-gravity + ~/<persona>-pc · :8790)"]
CORP["Level 1: Root 0 (Public Admin 1..2 | Enterprise 10 | Individual 90)\n(corporate-root-hub :8710 | workstation-control-tower :8790)"]
DEPT["Level 2: Statutory, Functional & Knowledge Nodes\n(domain-policy [5..9] | knowledge [20] | department [15..45])"]
SQUAD["Level 3: Operational Units, Tutors & Delivery Squads\n(squad [40..60] | personal tutor [86..89] · node-work-cockpit / tutor-study-cockpit)"]
PROJ["Level 4: Application, Case & Cloud Service Projects\n(project [80] · Contextual Human-Stack M1..M4 [85..95])"]
SUB <-->|"C-Backed Primitives &\nOSS-to-OSS Contributions"| OSS
OSS -->|"workstation-sync.sh\n(Coupled $HOME Runtime)"| APP
OSS <-->|"Downward: make sync\nUpward: Anonymized PR (/enrich)"| CORP
CORP <-->|"Downward: make delegate & make sync\nUpward: make complete & make promote"| DEPT
DEPT <-->|"Downward: make delegate & make sync\nHorizontal: make consult (rcap1)"| SQUAD
SQUAD <-->|"Downward: apply-project & make task\nUpward: make complete & /enrich"| PROJ
2.3.1. Summary of the 6 Orthogonal Stack & Layer Composition Patterns (entity_kind × layer_role × stack_groups)
Every workspace composes inherited human-stack / context-stack / framework-stack / plugin-stack packs (templates/*-stack/<pack>/) with its own sovereign node rules (templates/<layer-id>/rules/<layer-id>.md) along the 0..100 Lex Superior spectrum:
| Organizational / Epistemic Need | Who Owns the Workspace | Combined Stack Lineage (0..100 Lex Superior) |
How It Activates (1 Command) |
Effective Antigravity Agent Persona |
|---|---|---|---|---|
| 1. Corporate / Departmental Operations (Contracts, IP Registries, Ledgers, SOPs) | Root 0 (10) or Department (5..35) |
universal-base $\rightarrow$ business-base $\rightarrow$ corporate-operations + templates/<dept>/ |
zcg bootstrap-workspace / make sync |
Normative & SSoT Ledger Auditor (corporate-operations) |
| 2. Scientific / Statutory Knowledge Atlas (Axioms, Statutes, Geographic Registries) | Knowledge Node (gov-*, sci-*, geo-*, org-*, 20) |
universal-base $\rightarrow$ science-base $\rightarrow$ scientific-methodology + templates/<atlas>/ |
zcg bootstrap-workspace --role knowledge |
Epistemic Corpus Steward (E3..E0 & citation_receipt) |
3. Personal Academic Tutor (M1 -> M4) (Multi-specialty Socratic study hub) |
Individual Tutor Sub-Layer (<persona>-tutor, 86) |
universal-base $\rightarrow$ academic-base $\rightarrow$ academic-curriculum + tutor-study-cockpit |
zcg bootstrap-workspace --templates academic-base,academic-curriculum,tutor-study-cockpit |
Socratic M1..M4 Study Coach (specialty/* DMCP) |
4. Software & Cloud Engineering Squad (Microservices, GCP IaC, 100% isolated 1:1:1) |
Department (25) / Squad (50) / Project (80) |
Corp (10) $\rightarrow$ science-base $\rightarrow$ software-engineering $\rightarrow$ google-cloud-iac |
zcg apply-project <app> --parent-dir <dept> --template google-cloud-iac |
1:1:1 Software & Google Cloud IaC Architect |
5. Fiduciary SecOps & Legal Guardian (Cross-departmental NIST-01..04 & IP gate) |
dept-legal-ip (5) or dept-secops (5) |
science-base $\rightarrow$ software-engineering $\rightarrow$ nist-csf-security $\rightarrow$ secops-mcp-guardian |
zcg sync-layer --templates nist-csf-security,secops-mcp-guardian |
Zero-Trust Fiduciary & Ephemeral Bot Orchestrator |
6. Punctual Cross-Layer Consultation (Read-only rcap1 capsule across layers) |
Any consumer workspace | Active workspace + Ephemeral rcap1 Capsule (kr: true) |
make consult LAYER=dept-legal-ip QUERY="..." |
Zero-Residue Read-Only Inter-Layer Advisor |
2.3.2. From Zero to a Custom Governed Workspace in 3 Steps (Clone -> Open in Antigravity -> /sopen)
Once ~/zero-cloud-gravity is installed in $HOME, initializing or reconciling any repository (~/Projects/<your-repo>) requires zero manual boilerplate:
# A. First-Time Workspace Onboarding & Purpose Customization (State U0/U1 -> State U2 "configured"):
/sopen — Bootstrap and customize this workspace [<layer-id>] (role: [corporate|domain-policy|department|squad|knowledge|project], precedence: [1..95], templates: [science-base,software-engineering,google-cloud-iac,node-work-cockpit]):
1. Phase A (Bootstrap, Auto-Cockpit & Purpose Alignment): Run zcg bootstrap-workspace (activating the node's Cockpit at http://127.0.0.1:<8700+prec>/?env=dev), set our mission ([<1-sentence domain purpose>]), owner emails ([<owner@domain>,$(git config user.email)]), and Human-Stack ([domain-<area>:M1..M4]).
2. Phase B (Sovereign Rules & 1:1:1 Architecture): Tailor templates/<layer-id>/rules/<layer-id>.md with our domain invariants ([<RULE-01..03>]), replace all [Fill: ...] placeholders in README.md and .agents/LAYER_LINEAGE.md, and initialize docs/architecture/ and WORKTASK.md.
3. Phase C (Verification & Cross-Layer Consultation): Verify make consult against mounted knowledge nodes (gov-*/geo-*/sci-*/org-*) and confirm 0 drift via make sync && make diff.
4. Phase D (Configured Transition & Live Cockpit Ready): Mark "- [x] Complete Layer Setup Blueprint..." in WORKTASK.md (transitioning setup_state.state to "configured"), refresh the Node Cockpit (make activate-plugin), and output the initial git commit block.
# B. Ongoing 4-Step Upstream Upgrade & Version Reconciliation (State U2):
/sopen — Execute the 4-Step Upstream Upgrade & Version Reconciliation Flow:
1. Step 1 (Topology & Horizon Audit): Run make status, make horizon, and make upstream-diff.
2. Step 2 (Zero-Loss Reconciliation): Run make sync STRATEGY=reconcile (preserving templates/<layer-id>/rules/<layer-id>.md while syncing templates/*-stack/*).
3. Step 3 (Selective Stack & Plugin Review): Verify active templates/plugins and confirm make diff (0 drift).
4. Step 4 (Environment Gate & Closure): Validate make env-gate ENV=tst|pre|pro and close via /sclose.
3. Natural Code Evolution, Command Consciousness & Reproducible Metrics
Code structure, CLI commands, and repository health evolve deterministically across three coupled axes:
| Dimension / Lifecycle Point | Reproducible Target Metric / Invariant | Human & Architectural Value Added | Agent Deviation / Drift Signal (Validation Gate) |
|---|---|---|---|
1. Session Startup & Orientation (/sopen / make status) |
1 CLI call (< 0.2s) returning entity_identity, network_topology, work_orchestration, runtime_orchestration, setup_state, horizon, and agent_diff |
Eliminates 6–8 slow shell/git/file probes at session start and prevents blind brain/ scans |
Multi-step manual git/ls probing on startup or mutating read-only upstream templates |
2. Context-Window & Compaction (AGENTS.md & rcap1) |
<= 18,000 bytes global AGENTS.md (universal-base ~2.6 KB); 67%–88.7% token savings at L1/L0 (~16 ms) |
Zero rule eviction under Antigravity's 20,000-token inline budget; workspaces load only rules matching their active stack lineage |
AGENTS.md > 18 KB, workspace .agents/rules/ > 80 KB, or unindexed ##### headings |
3. Natural Code Maintainability & 1:1:1 Subpackages |
<= 384 lines source cap, <= 464 lines test/doc cap, <= 400 lines root README.md cap, $M_{\text{cluster}} < 1,200$, and 0 cross-peer _private imports |
Prevents God Components and coupled cluster drift; every module's 1:1:1 triad (Spec <-> Test <-> Code) is discoverable in $O(1)$ |
Source file > 384 lines, doc/test > 464 lines, cross-peer _private import, or missing Specification: header |
4. Bidirectional 1:1:1 Test Symbiosis (Unit -> Imp -> E2E) |
100.00% single-file isolated unit statement + branch coverage (74/74 modules), # Arrange / # Act / # Assert, 0-byte __init__.py |
Self-documenting executable specs where each test_<mod>.py reaches 100.00% statement and branch coverage in complete isolation |
Coverage < 100% in single-file isolation, missing standalone AAA markers ([missing-aaa]), or non-empty __init__.py |
5. Inter-Layer Governance & Stage Gates (dev -> tst -> pre -> pro) |
Strict 1-hop horizon, 4-role roster (owner, maintainer, validator, member) + consumer, and $N$-departmental surface consensus |
Unifies Business (Root 0 / Department SSoT ledgers & Lex Superior 0..100) and Technical (Squad / Project) delivery on feature/<TICKET-ID>-<slug> |
CROSS_TIER_PROMOTION_BYPASS, PATH_CONTAINMENT_VIOLATION, open P0 in tst, or unchecked WORKVALI.md items in pre |
ZeroCloudGravity MCP questions
What programming language and dependencies does it use?
It is built using 100% Python 3.14 Standard Library with zero external runtime dependencies.
Can it manage my Ubuntu workstation settings?
Yes, it can snapshot APT packages, Snaps, NPM globals, and GNOME settings into plain-text manifests and manage configuration files via GNU Stow.