ADK Runbooks#
This site serves as the central hub for documentation related to the Agent Development Kit (ADK) runbooks, a collection of automated procedures, security content, and operational guides designed to streamline cybersecurity operations. Here you’ll find information on setting up the project, configuring its components, understanding the multi-agent architecture, and leveraging the extensive Rules Bank for detection and response. Whether you’re a SOC analyst, detection engineer, or incident responder, this documentation aims to provide the necessary resources to effectively utilize and contribute to the ADK Runbooks ecosystem.
Project Documentation Contents#
This site contains detailed information about various components of the ADK Runbooks project. Explore the sections below to find specific documentation.
ADK Multi-Agent System Overview#
Details on the multi-agent architecture, project structure, configuration system, and how to run the example.
Setup & Configuration#
Instructions for setting up the project environment and configuring necessary components.
Note
The setup and configuration details below primarily pertain to the example multi-agent system provided in this repository.
Setup
Warning
Do NOT use uv
to run adk
with a pyproject.toml
file. (It causes intractable dependency resolution issues.)
Instead, do this:
git clone --recurse-submodules https://github.com/dandye/adk_runbooks.git
cd adk_runbooks/multi-agent
python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
adk run manager
Or with uv as pip replacement:
git clone --recurse-submodules https://github.com/dandye/adk_runbooks.git
cd adk_runbooks/multi-agent
python -m venv .venv
source .venv/bin/activate
uv pip sync requirements.txt
adk run manager
If you already cloned without submodules:
cd adk_runbooks
git submodule update --init --recursive
Configuration
There are two places to configure for your environment:
ADK Environment Variables
Copy
./multi-agent/manager/.env.example
to./multi-agent/manager/.env
Add your
GOOGLE_API_KEY
to the.env
file
MCP Security Tools
The MCP Security tools are included as a git submodule in
external/mcp-security/
Copy
./external/mcp-security/.env.example
to./external/mcp-security/.env
Configure your security tool API keys (Chronicle, SOAR, VirusTotal)
Core Documents#
These documents outline the foundational strategies, protocols, and plans for the ADK Runbooks project and the Rules Bank. They provide high-level guidance and operational frameworks.
Core Documents:
- Indicator Handling Protocols
- Detection Strategy Overview
- Project Plan: Enhance LLM Agent Context
- MCP Tool Best Practices & Usage Guide
- Analytical Query Patterns for AI Agents
- Automated Response Playbook Criteria
- Coding Conventions
- Data Normalization Map
- Detection Improvement Process for AI Agents
- Log Source Overview
- SOP & Automation Effectiveness Review Process
AI Documentation#
This section covers documents related to the design, operation, and review of AI systems within the security context.
Atomic Runbooks#
This section contains a collection of atomic runbooks, which are focused, reusable procedures for specific security tasks. These are typically categorized by the primary entity type they address (e.g., IP Address, Domain, Hash).
General Security Runbooks#
This area houses a broader collection of runbooks for various security operations, including comprehensive investigation guides, triage procedures, specific incident response plans, and detection engineering workflows.
General Security Runbooks:
- General Security Runbooks
- Advanced Threat Hunting (Hypothesis-Driven) Runbook
- Alert Investigation Summary Report Runbook
- Runbook: APT Threat Hunt
- Basic Endpoint Triage & Isolation Runbook
- Basic IOC Enrichment Runbook
- Case Event Timeline & Process Analysis Workflow
- Runbook: Generate Case Investigation Report
- Close duplicate/similar Cases Workflow
- Cloud Vulnerability Triage & Contextualization
- Common Investigation Steps
- Compare GTI Collection to IoCs, Events in SecOps
- Runbook: Create Investigation Report
- Runbook: Data Lake Queries
- Deep Dive IOC Analysis Runbook
- SOC Analyst Tier 2 Demo Runbook (SOAR Focus)
- Runbook: Detection-as-Code Workflow (Placeholder)
- Runbook: Generate Detection Report
- Detection Rule Validation & Tuning Runbook
- Group Cases Workflow
- Graphviz Dotfile
- Runbook: Group Cases v2
- Guided TTP Hunt Runbook (Example: Credential Access)
- Guidelines and Workflows
- Investigate a Case + external tools
- Investigate Google Threat Intelligence Collection ID (Enhanced)
- IOC Containment Runbook
- Runbook: IOC Threat Hunt
- Incident Response Plans (IRPs)
- Lateral Movement Detection Hunt (Example: PsExec/WMI)
- Malware Triage Runbook
- Runbook: Meta-Analysis (Placeholder)
- Post-Incident Review (PIR) Runbook
- Prioritize and Investigate a Case
- Proactive Threat Hunting based on GTI Campaign/Actor
- Suspicious Login Alert Triage Runbook
- Runbook: Alert Triage
- Runbook: UEBA Report Analysis
- Detection-as-Code Rule Tuning Workflow
Templates and Use Cases#
Here you’ll find templates to help standardize the creation of new runbooks and documentation detailing specific detection use cases.
Security Personas#
Understanding the roles and responsibilities of different security team members is crucial for effective collaboration and tailored procedures. This section describes various security personas.
Security Personas:
Agent Workflow References#
Comprehensive collection of agent workflow patterns and operational procedures.
MCP Tools Integration#
Documentation for the Model Control Protocol (MCP) tools integration and reference guides for various security platforms.
Development and Planning#
This section includes documents related to ongoing development, future planning, and suggestions for the ADK Runbooks project.
Development & Planning: