- Home
- MCP servers
- Weave Protocol
Weave Protocol
- typescript
0
GitHub Stars
typescript
Language
2 months ago
First Indexed
3 weeks ago
Catalog Refreshed
Documentation & install
Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.
Installation
Add the following to your MCP client configuration file.
Configuration
View docs{
"mcpServers": {
"tyox-all-weave_protocol": {
"command": "npx",
"args": [
"@weave_protocol/mund"
]
}
}
}You can deploy and use the Weave Protocol Security Suite MCP server to orchestrate secure AI agent interactions, enforce intent verification, provide threat scanning, and maintain auditable execution trails. This MCP server coordinates multiple protocols and exposes a REST API alongside local, script-driven execution, enabling you to build and operate secure, compliant AI ecosystems.
How to use
Install the MCP client and connect to the server components to start coordinating agents, enforcing policies, and verifying executions. You will register agents, submit tasks, and review compliance and audit trails. Use the unified orchestrator, council, and vault components to manage multi-agent workflows, security, and governance from a single integration point.
How to install
Prerequisites: You need Node.js and npm installed. You may also install Python and a modern shell if you plan to run additional tooling. Follow the steps below to set up the MCP server stack and start using it.
# 1) Install the MCP client package manager globally (example for npm users)
npm i -g npm
# 2) Install or prepare the MCP server components as needed by your environment
# The Weave Protocol packages are intended to be installed per component as shown in their package descriptions
# Mund for guardian scanning
npm i @weave_protocol/mund
# Hord for vaulting and sandboxing
npm i @weave_protocol/hord
# Domere for verification, orchestration & compliance
npm i @weave_protocol/domere
# Witan for governance and messaging
npm i @weave_protocol/witan
# API for a universal REST endpoint
npm i @weave_protocol/api
Optionally, you can run a bundled start flow if you have a project that exposes a single entry point for the MCP, or you can start individual services as needed in your environment.
Additional sections
Configuration and usage notes are provided to help you deploy, secure, and operate the MCP server stack in real-world environments. The following content covers how to use the main components, how to connect via the REST API, and how to monitor for compliance and recover from failures.
Mund (Guardian) manages secret and threat scanning. Hord (Vault) provides secure containment and sandboxing. Dōmere (Judge) handles intent verification, replay, and compliance. Witan (Council) enables consensus, communication, and governance.
To connect a client that wants to use the MCP server, you can use the REST API exposed by the Domere component via the unified API package. You can also interact with Mund for scanning and Hord for secure storage from your application code.
Security notes: enable encryption for secrets, apply access control policies, enable cryptographic audit trails, and anchor critical events to a blockchain layer for immutable records. Implement role-based access, regular secret rotation, and strict handoff verifications to minimize risk.
Troubleshooting
If a component fails to start, check logs for the Mund, Hord, Domere, and Witan services, verify environment variables, and ensure required dependencies are installed. Reconcile any drift detected in threads or tasks, and re-run the compliance checks after remediation.
Available tools
MundScanner
Real-time secret and threat detection from inputs and outputs across AI interactions.
HordVault
Encrypted secret storage with redaction and sandboxed execution capabilities.
ExecutionReplayManager
Forensic replay and integrity verification of actions within execution threads.
ThreadManager
Thread creation, tracking, and drift verification for intent-based tasks.
ComplianceManager
Automated SOC2/HIPAA/other compliance logging and reporting.
TaskScheduler
Priority-based task scheduling with dependencies, retries, and load balancing.
AgentRegistry
Agent lifecycle, heartbeat monitoring, and failover handling.
StateManager
Distributed shared state with locking, branching, and conflict resolution.
Orchestrator
Unified multi-agent coordination interface for task submission and registration.
ConsensusEngine
Voting and proposal finalization for multi-agent governance.