Configuration System Quick Reference

Configuration System Quick Reference#

File Locations#

multi-agent/manager/config/
├── agents/*.yaml              # Agent capabilities
├── mcp_tools_config.yaml      # MCP servers
├── tool_agent_mapping.yaml    # Delegation rules
└── config_loader.py           # Python loader

Agent Configuration Template#

agent_name: agent_id
display_name: "Human Readable Name"
description: "What this agent does"

expertise_areas:
  - area_1
  - area_2

mcp_servers:
  - name: toolset_name
    tools:
      - tool_1
      - tool_2

delegation_triggers:
  - "keyword phrase"
  - "another trigger"

capabilities:
  can_escalate_to: [agent_names]
  handles_alert_severity: [low, medium, high]

Common Delegation Patterns#

Request Type

Pattern

Delegates To

Alert Triage

triage.*alert

SOC Tier 1

SOAR Cases

SOAR.*case

SOC Tier 2

Malware Analysis

malware.*analysis

SOC Tier 3

Threat Research

threat.*actor|CTI

CTI Researcher

Incident Response

incident.*response

Incident Responder

Threat Hunting

threat.*hunt

Threat Hunter

Rule Creation

detection.*rule

Detection Engineer

Testing Commands#

# Validate configurations
cd multi-agent/manager/tests
python3 validate_configs.py

# Test delegation logic
python3 demo_delegation.py

# Run all tests
./run_tests.sh

Adding a New Agent#

  1. Create config/agents/new_agent.yaml

  2. Define expertise, tools, and triggers

  3. Update tool_agent_mapping.yaml if needed

  4. Test with demo_delegation.py

  5. No code changes required!

Debugging Delegation#

from config.config_loader import load_agent_config

config = load_agent_config()
agent = config.get_agent_for_request("your request here")
print(f"Selected: {agent}")

# Get all details
details = config.get_agent_capabilities(agent)
print(details)

Key Configuration Files#

mcp_tools_config.yaml#

  • MCP server paths

  • Tool categories

  • Environment settings

tool_agent_mapping.yaml#

  • Tool → Agent mappings

  • Request patterns

  • Expertise matrix

agents/*.yaml#

  • Individual agent capabilities

  • Tool permissions

  • Delegation triggers

Contents