Symbi is an AI-native agent framework for building autonomous, policy-aware agents that can safely collaborate with humans, other agents, and large language models. The Community edition provides core functionality with optional Enterprise features for advanced security, monitoring, and collaboration.
- Docker (recommended) or Rust 1.88+
- Qdrant vector database (for semantic search)
Using GitHub Container Registry (Recommended):
# Run unified symbi CLI
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
docker run --rm -it -v $(pwd):/workspace ghcr.io/thirdkeyai/symbi:latest bash# Build development environment
docker build -t symbi:latest .
docker run --rm -it -v $(pwd):/workspace symbi:latest bash
# Build the unified symbi binary
cargo build --release
# Test the components
cargo test
# Run example agents (from crates/runtime)
cd crates/runtime && cargo run --example basic_agent
cd crates/runtime && cargo run --example full_system
cd crates/runtime && cargo run --example rag_example
# Use the unified symbi CLI
cargo run -- dsl parse my_agent.dsl
cargo run -- mcp --port 8080
# Enable HTTP API (optional)
cd crates/runtime && cargo run --features http-api --example full_systemEnable RESTful HTTP API for external integration:
# Build with HTTP API feature
cargo build --features http-api
# Or add to Cargo.toml
[dependencies]
symbi-runtime = { version = "0.1.0", features = ["http-api"] }Key Endpoints:
GET /api/v1/health- Health check and system statusGET /api/v1/agents- List all active agentsPOST /api/v1/workflows/execute- Execute workflowsGET /api/v1/metrics- System metrics
symbi/
├── src/ # Unified symbi CLI binary
├── crates/ # Workspace crates
│ ├── dsl/ # Symbi DSL implementation
│ │ ├── src/ # Parser and library code
│ │ ├── tests/ # DSL test suite
│ │ └── tree-sitter-symbiont/ # Grammar definition
│ └── runtime/ # Agent Runtime System (Community)
│ ├── src/ # Core runtime components
│ ├── examples/ # Usage examples
│ └── tests/ # Integration tests
├── docs/ # Documentation
└── Cargo.toml # Workspace configuration
- DSL Grammar: Complete Tree-sitter grammar for agent definitions
- Agent Runtime: Task scheduling, resource management, lifecycle control
- Tier 1 Sandboxing: Docker containerized isolation for agent operations
- MCP Integration: Model Context Protocol client for external tools
- SchemaPin Security: Basic cryptographic tool verification
- RAG Engine: Retrieval-augmented generation with vector search
- Context Management: Persistent agent memory and knowledge storage
- Vector Database: Qdrant integration for semantic search
- Comprehensive Secrets Management: HashiCorp Vault integration with multiple auth methods
- Encrypted File Backend: AES-256-GCM encryption with OS keychain integration
- Secrets CLI Tools: Complete encrypt/decrypt/edit operations with audit trails
- HTTP API: Optional RESTful interface (feature-gated)
- Advanced Sandboxing: gVisor and Firecracker isolation (Enterprise)
- AI Tool Review: Automated security analysis workflow (Enterprise)
- Cryptographic Audit: Complete audit trails with Ed25519 signatures (Enterprise)
- Multi-Agent Communication: Encrypted inter-agent messaging (Enterprise)
- Real-time Monitoring: SLA metrics and performance dashboards (Enterprise)
- Professional Services and Support: Custom development and support (Enterprise)
Define intelligent agents with built-in policies and capabilities:
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");
}
}
}
Symbi provides enterprise-grade secrets management with multiple backend options:
- HashiCorp Vault: Production-ready secrets management with multiple authentication methods
- Token-based authentication
- Kubernetes service account authentication
- Encrypted Files: Local AES-256-GCM encrypted storage with OS keychain integration
- Agent Namespaces: Scoped secrets access per agent for isolation
# Encrypt secrets file
symbi secrets encrypt config.json --output config.enc
# Decrypt secrets file
symbi secrets decrypt config.enc --output config.json
# Edit encrypted secrets directly
symbi secrets edit config.enc
# Configure Vault backend
symbi secrets configure vault --endpoint https://vault.company.com- Complete audit trails for all secrets operations
- Cryptographic integrity verification
- Agent-scoped access controls
- Tamper-evident logging
- Tier 1 Isolation: Docker containerized agent execution
- Schema Verification: Cryptographic tool validation with SchemaPin
- Policy Engine: Basic resource access control
- Secrets Management: Vault integration and encrypted file storage
- Audit Logging: Operation tracking and compliance
- Enhanced Sandboxing: gVisor (Tier2) and Firecracker (Tier3) isolation (Enterprise)
- AI Security Review: Automated tool analysis and approval (Enterprise)
- Encrypted Communication: Secure inter-agent messaging (Enterprise)
- Comprehensive Audits: Cryptographic integrity guarantees (Enterprise)
# Run all tests
cargo test
# Run specific components
cd crates/dsl && cargo test # DSL parser
cd crates/runtime && cargo test # Runtime system
# Integration tests
cd crates/runtime && cargo test --test integration_tests
cd crates/runtime && cargo test --test rag_integration_tests
cd crates/runtime && cargo test --test mcp_client_tests- Getting Started - Installation and first steps
- DSL Guide - Complete language reference
- Runtime Architecture - System design
- Security Model - Security implementation
- API Reference - Complete API documentation
- Contributing - Development guidelines
crates/runtime/README.md- Runtime-specific docscrates/runtime/API_REFERENCE.md- Complete API referencecrates/dsl/README.md- DSL implementation details
Contributions welcome! Please see docs/contributing.md for guidelines.
Development Principles:
- Security first - all features must pass security review
- Zero trust - assume all inputs are potentially malicious
- Comprehensive testing - maintain high test coverage
- Clear documentation - document all features and APIs
- Secure code generation and refactoring
- Automated testing with policy compliance
- AI agent deployment with tool verification
- Knowledge management with semantic search
- Healthcare data processing with HIPAA compliance (Enterprise)
- Financial services with audit requirements (Enterprise)
- Government systems with security clearances (Enterprise)
- Legal document analysis with confidentiality (Enterprise)
Community Edition: MIT License
Enterprise Edition: Commercial license required
Contact ThirdKey for Enterprise licensing.
Symbi enables secure collaboration between AI agents and humans through intelligent policy enforcement, cryptographic verification, and comprehensive audit trails.
![@dependabot[bot]](https://avatars.githubusercontent.com/in/29110?s=64&v=4){kind=small}