Practices & Implementation Guides

The definitive guide to the AGENTS.md file, focusing on minimal, high-signal context for AI agents.

A guide on how to add multiple personas to an AGENTS.md file, with examples.

Executing automated verification using a Critic Agent to validate implementation artifacts against Spec contracts.

Step-by-step guide for creating, organizing, and maintaining Architecture Decision Records in your codebase.

Ultra-granular commit practice for agentic workflows, treating version control as reversible save points.

How to write Product Backlog Items that agents can read, execute, and verify—with templates and lifecycle guidance.

The practice of moving agent trajectories, tool results, and historical state from the context window to the filesystem to prevent cognitive overload and context rot.

Define agentic workflows in deterministic code rather than prompts to ensure reliability, type safety, and testable orchestration.

The practice of creating high-density Context Maps to guide agents through codebases and documentation.

A verification practice where a Critic Agent challenges the problem statement and assumptions before any specification or code is written.

Step-by-step guide for implementing Constitutional Review to validate code against both Spec and Constitution contracts.

Practical guide to creating and maintaining specs that evolve alongside your codebase.

The implementation phase where PBIs are executed against Specs, validated through quality gates, and integrated into the codebase.

How to create and maintain a Product Vision document that transmits taste to agents—inline in AGENTS.md or as a separate file.