Lib.rs

中文简体 | Español | Português | 日本語 | Deutsch

🚀 What is Symbiont?

Symbi is a Rust-native, zero-trust agent framework for building autonomous, policy-aware AI agents. It fixes the biggest flaws in existing frameworks like LangChain and AutoGPT by focusing on:

• Security-first: cryptographic audit trails, enforced policies, and sandboxing.
• Zero trust: all inputs are treated as untrusted by default.
• Enterprise-grade compliance: designed for regulated industries (HIPAA, SOC2, finance).

Symbiont agents collaborate safely with humans, tools, and LLMs — without sacrificing security or performance.

⚡ Why Symbiont?

🏁 Quick Start

Prerequisites

• Docker (recommended) or Rust 1.88+
• Qdrant vector database (for semantic search)

Run with Pre-Built Containers

# Parse an agent DSL file
docker run --rm -v $(pwd):/workspace ghcr.io/thirdkeyai/symbi:latest dsl parse /workspace/agent.dsl

# Run MCP Server
docker run --rm -p 8080:8080 ghcr.io/thirdkeyai/symbi:latest mcp

# Interactive development shell
docker run --rm -it -v $(pwd):/workspace ghcr.io/thirdkeyai/symbi:latest bash

Build from Source

# Build dev environment
docker build -t symbi:latest .
docker run --rm -it -v $(pwd):/workspace symbi:latest bash

# Build unified binary
cargo build --release

# Run REPL
cargo run -- repl

# Parse DSL & run MCP
cargo run -- dsl parse my_agent.dsl
cargo run -- mcp --port 8080

🔧 Key Features

• ✅ DSL Grammar – Define agents declaratively with built-in security policies.
• ✅ Agent Runtime – Task scheduling, resource management, and lifecycle control.
• 🔒 Sandboxing – Tier-1 Docker isolation for agent execution.
• 🔒 SchemaPin Security – Cryptographic verification of tools and schemas.
• 🔒 Secrets Management – HashiCorp Vault / OpenBao integration, AES-256-GCM encrypted storage.
• 📊 RAG Engine – Vector search (Qdrant) with hybrid semantic + keyword retrieval.
• 🧩 MCP Integration – Native support for Model Context Protocol tools.
• 📡 Optional HTTP API – Feature-gated REST interface for external integration.

📐 Symbiont DSL Example

metadata {
    version = "1.0.0"
    author = "Your Name"
    description = "Data analysis agent"
}

agent analyze_data(input: DataSet) -> Result {
    capabilities = ["data_analysis", "visualization"]
    
    policy data_privacy {
        allow: read(input) if input.anonymized == true
        deny: store(input) if input.contains_pii == true
        audit: all_operations
    }
    
    with memory = "persistent", requires = "approval" {
        if (llm_check_safety(input)) {
            result = analyze(input);
            return result;
        } else {
            return reject("Safety check failed");
        }
    }
}

🔒 Security Model

• Zero Trust – all agent inputs are untrusted by default.
• Sandboxed Execution – Docker-based containment for processes.
• Audit Logging – Cryptographically tamper-evident logs.
• Secrets Control – Vault/OpenBao backends, encrypted local storage, agent namespaces.

📚 Documentation

• Getting Started
• DSL Guide
• Runtime Architecture
• Security Model
• API Reference

🎯 Use Cases

• Development & Automation

  • Secure code generation & refactoring.
  • AI agent deployment with enforced policies.
  • Knowledge management with semantic search.

• Enterprise & Regulated Industries

  • Healthcare (HIPAA-compliant processing).
  • Finance (audit-ready workflows).
  • Government (classified context handling).
  • Legal (confidential document analysis).

📄 License

• Community Edition: MIT License
• Enterprise Edition: Commercial license required

Contact ThirdKey for enterprise licensing.

Symbiont enables secure collaboration between AI agents and humans through intelligent policy enforcement, cryptographic verification, and comprehensive audit trails.