Immaculate MCP
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@Immaculate MCP check whether workflow wf_42 may issue a $250 refund"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Arobi Immaculate MCP — Pre-Action AI Governance & Evidence Server
Official MCP Registry ID: io.github.Arobi-Technology-Alliance-Lab/immaculate · Glama Remote Connector: glama.ai/mcp/connectors/io.github.Arobi-Technology-Alliance-Lab/immaculate

Arobi Immaculate MCP is a self-contained, zero-dependency Model Context Protocol (MCP) server (stdio / Docker) built by Arobi Technology Alliance, Inc..
It executes 100% locally and offline using Node.js built-in node:crypto (Ed25519 signatures + canonical SHA-256 hash chains) so MCP clients can issue bounded route grants, evaluate pre-action authority checks (GRANTED, DENIED, or ESCALATED), select policy-justified models, and cryptographically verify tamper-evident evidence chains locally without an API key or external network calls.
Local Implementation:
src/engine.mjs+src/index.mjs(Node.js ≥ 18, zero external runtime dependencies, includesDockerfileandnpm test)Hosted Cloud Option (
Streamable HTTP):https://aura-genesis.org/api/v1/mcp(Documentation & 110 Free Credits)Published Benchmarks (SHA-256 verifiable): https://aura-genesis.org/benchmarks/immaculate/
Live Service Status: https://aura-genesis.org/status/
Included Tools (Local Execution)
Tool | Purpose |
| Seals WHO granted WHAT authority to WHICH workflow ( |
| Verifies grant signature, expiry, action allowlist, category caps, and consequence-tier auth freshness before execution — returning GRANTED, DENIED, or ESCALATED with a hash-chained |
| Classifies prompt complexity, filters capable candidates, picks the lowest-cost capable model or applies an explicit |
| Re-derives canonical SHA-256 |
Related MCP server: Living Stack MCP — Retired Beta Listing
Run Locally (Node.js ≥ 18)
git clone https://github.com/Arobi-Technology-Alliance-Lab/immaculate-mcp.git
cd immaculate-mcp
npm test
npm startStdio MCP Client Config (Claude Desktop, Cursor, Claude Code, Windsurf)
{
"mcpServers": {
"arobi-immaculate": {
"command": "node",
"args": ["/absolute/path/to/immaculate-mcp/src/index.mjs"]
}
}
}Run via Docker
docker build -t arobi-immaculate-mcp .
docker run -i --rm arobi-immaculate-mcpOptional Managed Cloud Endpoint (Streamable HTTP)
Teams that want persistent multi-tenant cloud storage, dashboards, and hosted Streamable HTTP access can also connect directly to Arobi's hosted connector (https://aura-genesis.org/api/v1/mcp, free tier starts with 110 credits at aura-genesis.org/immaculate/mcp):
{
"mcpServers": {
"arobi-immaculate-cloud": {
"url": "https://aura-genesis.org/api/v1/mcp",
"headers": {
"X-Arobi-Tenant-Key": "<your-arobi-tenant-key>"
}
}
}
}Verification, Scope Boundary & License
Arobi provides technical verification, policy enforcement, and tamper-evident evidence tooling. Arobi is not an insurer or a certification body; licensed insurers, regulators, and independent reviewers retain their own decisions.
Everything in this repository (src/engine.mjs, src/index.mjs, tests, Dockerfile, manifest, and documentation) is proprietary to Arobi Technology Alliance, Inc. under the Arobi Proprietary License (see LICENSE), as is the optional hosted service at https://aura-genesis.org/api/v1/mcp. Use is governed by the Arobi Terms of Service.
Available Tools
4 toolsarobi_authorizeAInspect
Decide whether an agent action is permitted against a sealed route grant, and seal a hash-chained AUTHORIZATION record. Returns GRANTED, DENIED, or ESCALATED with checks and reason code.
Inputs: action (required), principal (required), route_grant (required), consequence_tier (required), +3 more.
arobi_classify_and_routeAInspect
Classify a prompt and choose the cheapest route capable of the work from your candidate set, reporting whether any cost above the cheapest capable route is justified by required_capabilities or an explicit quality_exception policy. Set record=false to preview without appending to the local chain.
Inputs: prompt (required), candidates (required), record, quality_exception, +1 more.
arobi_route_grantBInspect
Seal WHO granted WHAT authority to WHICH workflow, with scope, category caps, attenuation, and expiry. Returns a locally Ed25519-signed route_grant and hash-chained ROUTE_GRANT record.
Inputs: scope (required), route_id (required), granted_by (required), delegated_to (required), +3 more.
arobi_verify_chainAInspect
Recompute a set of sealed records locally and report whether the SHA-256 hash chain and Ed25519 signatures are intact, pinpointing the first broken record index and reason if tampered.
Inputs: records (required).
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 4 tools
Each tool maps to a distinct step in the sealed-authorization lifecycle: grant creation, chain verification, action authorization, and prompt classification/routing. The only soft overlap is the word 'route' appearing in both arobi_route_grant and arobi_classify_and_route, but the descriptions clearly separate authority sealing from model/route selection.
All names use a consistent arobi_ prefix and snake_case, and three of four are verb-led (verify_chain, authorize, classify_and_route). arobi_route_grant is noun-first and less parallel, but the family pattern remains readable and predictable.
Four tools is thin for a system managing grants, authorizations, audit chains, and routing; each tool is heavyweight but there are no query/list or revocation companions. It sits at the borderline of under-scoped rather than clearly right-sized.
Creation and verification are covered (grant, authorize, verify_chain, classify_and_route), but there is no revocation or expiration-management tool, no way to list/inspect issued grants, and no key rotation or chain export. Agents can work around this only by treating the chain as append-only with no lifecycle control.
Maintenance
Related MCP Connectors
Bounded authorization and independently verifiable execution evidence for agents.
Read-only Remote MCP for externally grounded AI agent trust receipts.
Tamper-evident proof creation and verification for AI agents via MCP, A2A, and REST.
MCP-native Trust Infrastructure for AI Agents. Persistent encrypted memory with Trust Quotient.
Related MCP Servers
- AlicenseNot gradedqualityCmaintenanceEnables AI agents to operate through a local, signed boundary that blocks prompt injection and secret leakage, verifies outputs, preserves cross-session memory, and provides offline-verifiable receipts. It also exposes 900+ MCP tools for discoverable agent actions.7MIT
- AlicenseAqualityBmaintenanceRetired snapshot of the former 15-tool Apache-2.0 beta. Use Living Stack Community for the free seven-tool proof edition; Complete Local is the paid runtime with memory, recovery, signed traces, release verification, and multi-agent workflows.14Apache 2.0
- AlicenseAqualityBmaintenanceFree seven-tool proof loop for MCP agents: scoped sessions, bounded authorization, typed outcomes, evidence-bound claims, and tamper-evident local closure. Upgrade to Complete Local for durable memory, recovery, signed traces, release verification, and multi-agent workflows.7Apache 2.0
- AlicenseBqualityBmaintenanceEnables software agents to maintain a signed, evidence-aware local ledger of claims, with policy-gated proposals, conflict queries, and constitution-pinned access through an MCP stdio server.73Apache 2.0