Technical documentation

The theory behind the kit

agent-ready is a small, context-free set of conventions for working with AI agents on code. These two deep-dives are the reasoning underneath — written for developers and software architects standing up serious AI collaboration, at maximum technical depth.

Not the getting-started guide — that is the README. This is the why.
01

The Coordination Model

Multi-agent collaboration on a shared codebase, framed as a distributed-systems problem: identity under a shared credential, durable vs. volatile channels, why green-at-the-branch is not green-at-the-merge, the single-writer serialization point, at-least-once delivery, and the economics that decide whether you need any of it. → read

02

Architecture

The architecture of agent context: context as an interface, a single source of truth with thin per-tool adapters, layered composition and precedence, the durable memory tier, the testability pattern, the security checkpoint, and the design rationale that ties it together. → read