Skip to main content

Base agent class for all Beast Mode agents

Project description

Beast Agent

Base agent class for ALL Beast Mode agents

PyPI version Python Versions License: MIT Code style: black


🎯 Purpose

The foundational base class (BaseAgent) that EVERY Beast Mode agent inherits from.

This package provides:

  • ✅ Standardized agent lifecycle (startup, shutdown, health checks)
  • ✅ Message handling via beast-mailbox-core integration
  • ✅ Agent registration and discovery
  • ✅ Capability declaration and management
  • ✅ Optional logging/telemetry hooks
  • ✅ Configuration management

Architectural Clarity:

  • beast-agent = Low-level base class (every agent IS-A beast-agent)
  • beast-agentic-framework = High-level orchestration (multi-agent coordination)

🚀 Quick Start

Installation

pip install beast-agent

Configuration

Configure Redis connection via environment variables:

For authenticated clusters (recommended for production):

export REDIS_HOST="your-redis-host"
export REDIS_PORT="6379"
export REDIS_PASSWORD="your-password"
export REDIS_DB="0"  # Optional, defaults to 0

For unauthenticated connections:

export REDIS_URL="redis://localhost:6379"

Or pass directly to constructor:

  • None (default): Automatically reads REDIS_HOST, REDIS_PORT, REDIS_PASSWORD, REDIS_DB from environment variables (supports authentication), or falls back to REDIS_URL if not set
  • String URL: mailbox_url="redis://localhost:6379" (unauthenticated only)
  • MailboxConfig object: For explicit configuration (see Authentication section)

Create Your First Agent

from beast_agent import BaseAgent
from beast_agent.decorators import capability

class MyAgent(BaseAgent):
    """Simple agent with a single capability"""
    
    def __init__(self):
        super().__init__(
            agent_id="my-agent",
            capabilities=["process_data"]
        )
    
    @capability("process_data")
    async def process_data(self, data: dict) -> dict:
        """Process data and return results"""
        # Your agent logic here
        return {"status": "processed", "result": data}

# Run the agent
async def main():
    agent = MyAgent()
    await agent.startup()
    # Agent is now ready to handle messages
    
if __name__ == "__main__":
    import asyncio
    asyncio.run(main())

🔐 Connecting with Authentication

If your Redis cluster requires authentication, use MailboxConfig instead of a URL string:

from beast_agent import BaseAgent
from beast_mailbox_core import MailboxConfig
import os

class AuthenticatedAgent(BaseAgent):
    """Agent that connects to authenticated Redis cluster."""
    
    def __init__(self):
        # Create MailboxConfig with password
        mailbox_config = MailboxConfig(
            host=os.getenv("REDIS_HOST", "localhost"),
            port=int(os.getenv("REDIS_PORT", "6379")),
            password=os.getenv("REDIS_PASSWORD"),  # Required for authenticated clusters
            db=0
        )
        
        super().__init__(
            agent_id="authenticated-agent",
            capabilities=["example"],
            mailbox_url=mailbox_config  # Pass MailboxConfig object, not URL string
        )
    
    async def on_startup(self) -> None:
        self._logger.info("Connected to authenticated cluster!")
    
    async def on_shutdown(self) -> None:
        self._logger.info("Disconnecting...")

Note: The mailbox_url parameter accepts:

  • String URL: "redis://localhost:6379" (for unauthenticated connections)
  • MailboxConfig object: For authenticated or advanced configurations (recommended for production)
  • None (default): Automatically reads REDIS_HOST, REDIS_PORT, REDIS_PASSWORD, REDIS_DB from environment variables (supports authentication), or falls back to REDIS_URL if not set

For production clusters with authentication, simply set REDIS_HOST, REDIS_PORT, REDIS_PASSWORD environment variables and use mailbox_url=None (the default).

See examples/authenticated_agent.py for a complete example.


🔍 Cluster Discovery

After your agent starts, it automatically registers on the cluster. Other agents can discover you, and you can discover them:

Discover All Agents

# After startup, discover all active agents on the cluster
await agent.startup()

# List all agent IDs
all_agents = await agent.discover_agents()
print(f"Found {len(all_agents)} agents: {all_agents}")

Get Agent Metadata

# Get metadata for a specific agent
agent_info = await agent.get_agent_info("other-agent-id")
if agent_info:
    print(f"Agent ID: {agent_info['agent_id']}")
    print(f"Capabilities: {agent_info['capabilities']}")
    print(f"State: {agent_info['state']}")
    print(f"Registered: {agent_info['registered_at']}")

Find Agents by Capability

# Find all agents that provide a specific capability
search_agents = await agent.find_agents_by_capability("search")
for agent_info in search_agents:
    print(f"Found search agent: {agent_info['agent_id']}")
    
    # Send message to discovered agent
    await agent.send_message(
        target=agent_info["agent_id"],
        message_type="HELP_REQUEST",
        content={"request": "I need help with search"}
    )

Redis Keys Used

The cluster uses these Redis keys for discovery:

  • beast:agents:all - Set containing all active agent IDs
  • beast:agents:{agent_id} - Hash with agent metadata (expires after 60 seconds)

Complete Discovery Example

import asyncio
from beast_agent import BaseAgent
from beast_mailbox_core import MailboxConfig

class DiscoveryAgent(BaseAgent):
    def __init__(self):
        mailbox_config = MailboxConfig(
            host="your-redis-host",
            port=6379,
            password="your-password",
            db=0
        )
        super().__init__(
            agent_id="discovery-agent",
            capabilities=["discover", "communicate"],
            mailbox_url=mailbox_config
        )
    
    async def on_startup(self) -> None:
        self.register_handler("HELP_REQUEST", self.handle_help)
        
        # Discover all agents on cluster
        all_agents = await self.discover_agents()
        self._logger.info(f"Found {len(all_agents)} agents: {all_agents}")
        
        # Find agents with specific capability
        helpers = await self.find_agents_by_capability("help")
        for helper in helpers:
            self._logger.info(f"Found helper: {helper['agent_id']}")
    
    async def handle_help(self, content: dict) -> None:
        sender = content.get("sender")
        self._logger.info(f"Received help request from {sender}")

async def main():
    agent = DiscoveryAgent()
    await agent.startup()
    
    # Keep running
    await asyncio.sleep(3600)
    
    await agent.shutdown()

if __name__ == "__main__":
    asyncio.run(main())

📋 Agent Patterns

Per-Repo Agent Pattern

class RepoAgent(BaseAgent):
    """Agent that monitors a single repository"""
    
    def __init__(self, repo_name: str):
        super().__init__(
            agent_id=f"repo-agent-{repo_name}",
            capabilities=["code_review", "pr_validation", "security_scan"]
        )
        self.repo_name = repo_name
    
    @capability("code_review")
    async def review_code(self, pr_number: int) -> dict:
        """Review code in PR"""
        # Review logic
        return {"status": "reviewed", "pr": pr_number}

Per-Branch Agent Pattern

class BranchAgent(BaseAgent):
    """Agent specific to a branch"""
    
    def __init__(self, repo_name: str, branch_name: str):
        super().__init__(
            agent_id=f"branch-agent-{repo_name}-{branch_name}",
            capabilities=["deployment", "testing", "monitoring"]
        )
        self.repo = repo_name
        self.branch = branch_name
    
    @capability("deployment")
    async def deploy(self, environment: str) -> dict:
        """Deploy branch to environment"""
        # Deployment logic
        return {"status": "deployed", "environment": environment}

🔧 Features

Agent Lifecycle

agent = MyAgent()
await agent.startup()  # Initialize and register
# ... agent is running ...
await agent.shutdown()  # Graceful cleanup

Message Handling

# Register handler for message type
agent.register_handler("TASK_REQUEST", handle_task_request)

# Send message to another agent
await agent.send_message(
    target="other-agent-id",
    message_type="HELP_REQUEST",
    content={"task": "analyze_code"}
)

Health Checks

health = agent.health_check()
print(f"Healthy: {health.healthy}")
print(f"State: {health.state}")

📚 Documentation


🧪 Testing

# Run all tests
pytest

# Run with coverage
pytest --cov=src/beast_agent --cov-report=html

# Run specific test
pytest tests/test_base_agent.py

# Run integration tests (requires Redis and beast-mailbox-core)
# Note: Integration tests are automatically skipped if dependencies unavailable
pytest tests/test_mailbox_integration.py

# Run all tests including integration (if Redis available)
pytest tests/

Integration Testing

Integration tests require:

  • Redis: Running locally (via Docker) or in CI (via service containers)
  • beast-mailbox-core: Installed as dependency

Local Testing:

  • Redis Docker container is automatically managed via conftest.py fixtures
  • Tests gracefully skip if Redis/Docker unavailable
  • Uses separate test database (db=15) to avoid conflicts

CI Testing:

  • Redis service container automatically provided in GitHub Actions
  • Integration tests run against real Redis in CI

🤝 Integration

Required Dependencies

  • beast-mailbox-core >= 0.3.0 - Messaging and discovery

Optional Dependencies

  • beast-observability - Enhanced telemetry
  • Any cloud platform (works with AWS, GCP, Azure, on-prem)

📦 Package Status

Tier: 1 (Foundation)
Phase: Development
Coverage: Target 90%+
Quality: Target zero defects


🔗 Related Packages

  • beast-mailbox-core - Redis-backed mailbox utilities
  • beast-agentic-framework - Multi-agent orchestration
  • beast-observability - Unified telemetry
  • beast-redaction-client - Data classification

📄 License

MIT License - see LICENSE for details.


🙏 Acknowledgments

Part of the Beast Mode multi-agent framework ecosystem.


Built with ❤️ by the Beast Mode team

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

beast_agent-0.1.4.tar.gz (27.4 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

beast_agent-0.1.4-py3-none-any.whl (14.3 kB view details)

Uploaded Python 3

File details

Details for the file beast_agent-0.1.4.tar.gz.

File metadata

  • Download URL: beast_agent-0.1.4.tar.gz
  • Upload date:
  • Size: 27.4 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/6.1.0 CPython/3.13.7

File hashes

Hashes for beast_agent-0.1.4.tar.gz
Algorithm Hash digest
SHA256 7ebc6cc3ad1bab178aca592bce8a567c963114517eb60dd51ac9c083f659d6f6
MD5 da776f0bde6bb4165b99b1f047950f6c
BLAKE2b-256 42b6d950153f91b5df82bebb052c6116cd1a4daf27c205440a82d4c0268762f1

See more details on using hashes here.

Provenance

The following attestation bundles were made for beast_agent-0.1.4.tar.gz:

Publisher: publish.yml on nkllon/beast-agent

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file beast_agent-0.1.4-py3-none-any.whl.

File metadata

  • Download URL: beast_agent-0.1.4-py3-none-any.whl
  • Upload date:
  • Size: 14.3 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/6.1.0 CPython/3.13.7

File hashes

Hashes for beast_agent-0.1.4-py3-none-any.whl
Algorithm Hash digest
SHA256 9daf88f2d79e968f571612ebee9e153b10b5ab4b508a79cecbc99b79e65fd579
MD5 0f14c01af8beb134707cd11d018163ac
BLAKE2b-256 3af1b18d541015dac072b61bb997504679e939252000a06130eea8e93ff8f89e

See more details on using hashes here.

Provenance

The following attestation bundles were made for beast_agent-0.1.4-py3-none-any.whl:

Publisher: publish.yml on nkllon/beast-agent

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page