Skillget

AMS

mcp
Security Audit
Pass
Health Pass
  • License — License: NOASSERTION
  • Description — Repository has a description
  • Active repo — Last push 0 days ago
  • Community trust — 11 GitHub stars
Code Pass
  • Code scan — Scanned 12 files during light audit, no dangerous patterns found
Permissions Pass
  • Permissions — No dangerous permissions requested
Purpose
This polyglot platform acts as a durable orchestration and memory layer for MCP clients. It provides extensive task management capabilities, AI model routing, and peer-to-peer consensus mechanisms.

Security Assessment
Overall Risk: Medium. The code requires an `.env` file for configuration, which is the standard and safe method for handling environment variables (no hardcoded secrets were detected by the scan). However, the system inherently processes and orchestrates data across external networks. It makes network requests to multiple AI providers (Anthropic, OpenAI, Moonshot) and supports distributed P2P consensus, meaning it communicates heavily outside your local machine. No dangerous execution patterns or excessive permissions were found in the light scan, but the broad integration scope requires proper network and access controls.

Quality Assessment
The project is actively maintained, with its most recent code push happening today. It features a very well-documented ecosystem, including detailed operational and security guidelines. The repository has accrued 11 GitHub stars, indicating a small but existent user base. Note that while the automated health check could not automatically resolve the license type, the manual README states it uses "Apache-2.0 WITH Commons-Clause", which restricts commercial use.

Verdict
Use with caution due to heavy external network routing and a non-standard commercial use restriction on the license.
SUMMARY

Audit Memory System (AMS): Durable orchestration and memory layer for MCP clients.

README.md

Colibri

Polyglot (Node.js + Python) platform for task orchestration, multi-model AI routing, and peer-to-peer consensus.

What it does

  • 484 MCP tools for task management, audit trails, agent orchestration, and proof generation
  • Multi-model router across 3 AI providers (Anthropic, OpenAI, Moonshot) with weighted scoring
  • P2P consensus with Byzantine fault tolerance, state forking, and deterministic rule execution
  • Obsidian plugin for bidirectional vault sync with CRDT conflict resolution

Quick start

npm install
cp .env.example .env   # Edit with your config
npm start              # Start MCP server

See docs/guides/quickstart.md for the full walkthrough.

Documentation

All docs are in docs/:

Section Content
Concepts Core features named α through ν (Greek alphabet)
Specs 17 locked protocol specifications
Architecture Component map, data model, system design
Reference CLI, REST API, MCP tools, config, DSL
Guides Installation, quickstart, how-to guides
Operations Deploy, monitor, backup, troubleshoot
Security Auth, hardening
Developers Contributing, testing, architecture gate

Stack

  • Node.js (ESM) — MCP server, HTTP API, middleware, domains, dashboard
  • Python (FastAPI) — BFT consensus, rule engine, fork protocol, gossip
  • SQLite — single-writer database, 22 tables
  • Next.js + React — web UI

Commands

npm start          # Start server
npm test           # Run tests
npm run lint       # Lint check
npm run dev        # Dev mode with file watching
npm run dashboard  # Web dashboard
npm run db:check   # Database integrity
npm run arch:gate  # Architecture validation

License

Apache-2.0 WITH Commons-Clause

Reviews (0)

No results found