RiskLab
Getting Started
Installation
Requirements
Install from Source
LLM Configuration
Verify Installation
Architecture Overview
Components
Execution Flow
Cyclic vs. Acyclic Mode
Registry Pattern
Quickstart
1. Prepare
2. Run
3. Check Results
4. Write Your Own Config
Core Concepts
Risk Taxonomy
Reference Detectors (Included)
Full 13-Risk Taxonomy
Interaction Categories
Key Abstractions
CommunicationTopology
InformationFlowConfig
TaskConfig
Risk Detectors
Core Modules
LLM Configuration
Provider Setup
API Key Resolution
Provider Auto-Detection
Using the Client
Per-Agent Models
Topology
CommunicationTopology
TimeVaryingTopology
InformationFlowConfig
YAML Configuration
Agents
LLMAgent
EnhancedLLMAgent
MarketSellerAgent
YAML Configuration
Agent Registry
Environments
Core Interface
Built-in Environments
YAML Configuration
Protocols
Core Interface
Built-in Protocols
YAML Configuration
Choosing a Protocol
Risks & Evaluation
Risk Detectors
Built-in Detectors
YAML Configuration
Risk Registry
Evaluation
Trajectory
Trajectory Logger
Metrics
Task Evaluator
User Guides
Running Experiments
Using Reproducible Example Scripts
Using the Python API
Multi-Seed Runs
Output Structure
Inspecting Configs
Experiment Configuration
Top-Level Structure
Experiment
Topology
Agents
Environment
Protocol
Task
Risks
Complete Example
Extending the Framework
Custom Agent
Custom Environment
Custom Risk Detector
Custom Protocol
Custom Metric
MCP & Skills
Integration Checklist
Tutorials
Risk I: Tacit Collusion
What is Tacit Collusion? (Risk 2 / R2)
Prerequisites
Step 1 — Configure Your API Key
Step 2 — Understand the Config
Step 3 — Run the Experiment
Step 4 — Understand the Execution Flow
Step 5 — Interpret the Results
Step 6 — Customize and Extend
Troubleshooting
What’s Next?
Risk II: Strategic Misreporting
What is Strategic Misreporting? (Risk 9 / R9)
Prerequisites
Step 1 — Configure Your API Key
Step 2 — Understand the Config
Step 3 — Run the Experiment
Step 4 — Understand the Execution Flow
Step 5 — Interpret the Results
Step 6 — Customize and Extend
Troubleshooting
What’s Next?
Risk III: Normative Deadlock
What is Normative Deadlock? (Risk 10 / R10)
Prerequisites
Step 1 — Configure Your API Key
Step 2 — Understand the Config
Step 3 — Run the Experiment
Step 4 — Understand the Execution Flow
Step 5 — Interpret the Results
Step 6 — Customize and Extend
Troubleshooting
What’s Next?
Risk IV: Excessive Rigidity
What is Excessive Rigidity? (Risk 13 / R13)
Prerequisites
Step 1 — Configure Your API Key
Step 2 — Understand the Config
Step 3 — Run the Experiment
Step 4 — Understand the Execution Flow
Step 5 — Interpret the Results
Step 6 — Customize and Extend
R2 vs R13 — Key Differences
Troubleshooting
What’s Next?
Additional Resources
CLI Reference
Config Inspector
Flags
Examples
Sample Output
Citation
License
Links
RiskLab
Index
Index