create-agent-adapter

Create Sprint agent adapters with server, UI, CLI interfaces and four-exports convention.

Updated Mar 25, 2026
One-click install
npx skills add https://github.com/gancim/sprint --skill create-agent-adapter-gancim
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: create-agent-adapter
Source: https://github.com/gancim/sprint/tree/main/.agents/skills/create-agent-adapter
Command: npx skills add https://github.com/gancim/sprint --skill create-agent-adapter-gancim

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve?

Technical guide for creating a new Sprint agent adapter. Use when building a new adapter package, adding support for a new AI coding tool (e.g. a new CLI agent, API-based agent, or custom process), or when modifying the adapter system. It covers the required interfaces, module structure, registration points, and conventions derived from the existing claude-local and codex-local adapters.

Core Features & Use Cases

  • Provides a repeatable blueprint for implementing the server, UI, and CLI interfaces and for exporting the adapter module in the four-exports convention.
  • Documents how to implement execute, testEnvironment, parse, and configuration builders, plus how to wire the adapter into server, UI, and CLI registries.
  • Describes the lifecycle of an adapter package, including agentConfigurationDoc, models, and registration in the registries.

Quick Start

Create a new adapter package under packages/adapters/<name>, implement the four exports, and wire it into the server, UI, and CLI registries.

Frequently Asked Questions about create-agent-adapter

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I create a Sprint adapter for a custom AI agent?

To create a Sprint adapter for a custom AI agent, implement the server, UI, and CLI interfaces under packages/adapters/<name> using the four-exports convention. The adapter must bridge Sprint's orchestration layer to your specific AI runtime by providing execute, testEnvironment, parse, and configuration builders.

What is the four-exports convention for Sprint agent adapters?

The four-exports convention for Sprint agent adapters requires packaging the adapter module under packages/adapters/<name> with specific exports that implement server, UI, and CLI interfaces. This structure includes a server-side session codec, agent configuration documentation, and model metadata.

How do I wire a new AI coding tool into Sprint server, UI, and CLI registries?

You wire a new AI coding tool into Sprint registries by implementing the required adapter interfaces and registering them across the server, UI, and CLI layers. The adapter must implement execute, testEnvironment, parse, and configuration builders to connect the orchestration layer to the AI runtime.

Do I need to implement environment handling when building a Sprint adapter?

Yes, you need to implement robust environment handling when building a Sprint adapter. The adapter guidelines require a testEnvironment builder and a server-side session codec to ensure the AI runtime operates correctly within the orchestration layer.

Can I use the Sprint adapter blueprint for both CLI and API-based agents?

Yes, you can use the Sprint adapter blueprint for CLI agents, API-based agents, and custom processes. The blueprint provides a repeatable structure for implementing the server, UI, and CLI interfaces regardless of the specific AI runtime being bridged.