Xcapit Labs / ArgenTor
Secure Multi-Agent AI Orchestration in Rust
Build production-grade AI agent systems with defense-in-depth security. WASM-sandboxed plugins, centralized MCP proxy, human-in-the-loop approval, and built-in compliance for GDPR, ISO 27001, ISO 42001, and DPGA.
Capabilities
What ArgenTor Does
WASM-Sandboxed Plugins
Every agent plugin runs in an isolated WASM sandbox via wasmtime. Memory limits, syscall filtering, and capability-based permissions prevent any plugin from affecting the host system.
Multi-Agent Orchestration
Coordinate multiple AI agents through typed message channels, broadcast/unicast routing, and priority-based scheduling. 5 communication channels with backpressure and deadlock detection.
MCP Centralized Proxy
Single point of control for all Model Context Protocol tool calls. Rate limiting, cost tracking, audit logging, and policy enforcement across all agent-to-tool interactions.
Human-in-the-Loop Approval
Configurable approval workflows for high-risk operations. Agents can request human review before executing sensitive actions — with timeout policies and escalation rules.
Compliance Built-In
Native support for GDPR data handling, ISO 27001 information security, ISO 42001 AI management, and DPGA digital public good standards. Compliance is architecture, not an afterthought.
Multi-Provider LLM Support
Connect to any LLM provider — OpenAI, Anthropic, local models. Automatic fallback, load balancing, and cost optimization across providers with unified API.
Quality
Engineering Excellence
Zero Clippy Warnings
The entire codebase — 13 crates — compiles with zero Clippy warnings. Strict linting enforces Rust best practices and catches potential issues at compile time.
483 Passing Tests
Comprehensive test coverage spanning unit tests, integration tests, and end-to-end scenarios. Every component is tested in isolation and in composition.
Defense-in-Depth Security
Multiple security layers: WASM sandbox isolation, capability-based permissions, encrypted state management, rate limiting, and cryptographic audit trails.
DPGA Compliant
Designed to meet Digital Public Goods Alliance standards — open-source, privacy-preserving, and built for international development contexts.
Our Journey
Born from Real-World AI Agent Challenges
ArgenTor emerged from building AI agent systems for enterprise clients who needed security guarantees that existing frameworks couldn't provide.
The Problem
Building AI agents for enterprise clients, we found that existing Python-based frameworks offered no real security boundaries. Any agent could access any resource, making compliance impossible.
Architecture in Rust
Chose Rust for memory safety guarantees and WASM ecosystem maturity. Designed the 13-crate architecture with clear separation between orchestration, sandboxing, communication, and compliance.
MCP & Compliance
Integrated Model Context Protocol for standardized tool access. Built compliance modules for GDPR, ISO 27001, ISO 42001, and DPGA with automated policy enforcement.
Open Source & Enterprise
Released as open source with enterprise support. Production deployments for enterprise automation, development workflows, and compliance-heavy industries.
Systems-Level Architecture
ArgenTor leverages Rust's safety guarantees and WASM's isolation model for enterprise-grade AI orchestration.
Zero-cost abstractions, memory safety without garbage collection, and Tokio async runtime for high-concurrency agent orchestration.
WebAssembly sandboxing via wasmtime with configurable memory limits, fuel metering, and capability-based permission model.
Model Context Protocol for standardized agent-to-tool communication. Centralized proxy with rate limiting, audit logging, and policy enforcement.
Roadmap
Vision 2026
ArgenTor is becoming the standard for secure, compliant AI agent orchestration in enterprise and government.
Use Cases
Who Uses ArgenTor
Enterprise Automation
Companies deploy multi-agent systems for document processing, customer service, and internal workflow automation — with security boundaries that satisfy compliance teams.
Development Workflows
Development teams use ArgenTor to orchestrate AI coding agents, review bots, and CI/CD automation with fine-grained access control and human-in-the-loop approvals.
Compliance-Heavy Industries
Finance, healthcare, and government organizations run AI agents with built-in compliance for GDPR, ISO 27001, and sector-specific regulations.
FAQ
Frequently Asked Questions
Why Rust instead of Python?
Rust provides memory safety guarantees without garbage collection, making it ideal for security-critical AI infrastructure. WASM sandboxing, compile-time error prevention, and zero-cost abstractions mean ArgenTor agents run faster and more safely than Python alternatives.
What is WASM sandboxing?
Each agent plugin is compiled to WebAssembly and runs in an isolated sandbox. The sandbox enforces memory limits, restricts system calls, and uses capability-based permissions. A misbehaving plugin cannot affect the host or other plugins.
How does human-in-the-loop work?
You configure approval policies that define which actions require human review. When an agent attempts a high-risk operation, it pauses and sends a review request. Humans can approve, deny, or modify the action. Timeout policies ensure the system doesn't hang indefinitely.
Is ArgenTor compatible with existing MCP tools?
Yes. ArgenTor implements the full Model Context Protocol specification. Any MCP-compatible tool server works with ArgenTor's centralized proxy, which adds rate limiting, cost tracking, and policy enforcement on top.
Ready to build secure AI agent systems?
Whether you need enterprise automation, development workflows, or compliance-grade AI orchestration — ArgenTor provides the secure foundation.