Agents
Agents Overview
Choose a track for building, evaluating, and debugging agents.
End-to-end workflows
Workflow-first guides: path selection, agent types, and practical runbooks.
Runtime setup
Register external or sandbox agents, wire SDK wrappers, and develop locally.
Coding agents
Run coding CLIs in seeded workspaces with scenarios, harness config, and grading.
Simulation and tools
Define tools, MCP integration, ledger schema, and task seeds.
Interaction patterns
Model multi-agent topologies and multi-turn session behavior.
Observability
Inspect traces, forward live events, and troubleshoot failure classes.
Reference examples
Framework-specific wrappers: OpenAI, Anthropic, Strands, and barebones Python.