Guide
How to build a reliable AI agent
A practical path from failure modes to structured context and decisions your agent can use.
Read the guideFor agent builders
Resources
Guides, research, and engineering for the agents you build.
Guides and engineering
9 guides
Engineering
Sandbox safety + a decision layer
Combine isolated execution with business rules for coding agents that act on the right change.
Read the guide
Technical
Why specialize a reasoning model?
Understand what changes when reasoning is shaped around a business domain.
Read the guide
Technical
What context graphs actually solve
Connect the facts, relationships and sources that make context useful to an agent.
Read the guide
Technical
Learning that carries forward
How feedback, evaluation and stable trajectories can build skill across runs.
Read the guide
Architecture
Keep decisions when agents change
Why agent continuity needs portable business logic, beyond switching the underlying model.
Read the guide
Engineering
More than permission to call a tool
Use evidence, process and validation to decide whether a specific action should execute.
Read the guide
Engineering
Working with multiple coding agents
Where parallel agents lose alignment, and how to keep a product coherent as they work.
Read the guide
Guide
What goes wrong in production
Recognize failure modes before your agent makes commitments or changes a business system.
Read the guideNo matching resources.
Try another term or explore the full library.
Keep exploring