swe-af

Coordinate autonomous AI agents to plan, code, and verify software builds.

969|165|Updated Feb 4, 2026
One-click install
npx skills add https://github.com/Agent-Field/SWE-AF --skill swe-af-agent-field
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: swe-af
Source: https://github.com/Agent-Field/SWE-AF/tree/main/docs
Command: npx skills add https://github.com/Agent-Field/SWE-AF --skill swe-af-agent-field

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Coordinating a fleet of AI agents to scope, design, implement, review, and verify software across repositories can be error-prone and slow when done manually. SWE-AF automates planning, coding, merging, and validation to deliver end-to-end software builds from a single natural-language goal.

Core Features & Use Cases

  • One-call automation: Trigger planning, coding, reviewing, merging, and verification from a single API call.
  • Multi-repo orchestration: Manage changes across multiple repositories with isolated work trees and safe merges.
  • Resilient execution: Durable checkpoints, runtime plan mutation, and graceful degradation with explicit debt tracking.

Quick Start

Execute a full SWE-AF build by posting a plan request to the control plane with your goal and repository details.

Frequently Asked Questions about swe-af

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

FAQPage Schema
How do I coordinate autonomous AI agents to build software across multiple repositories?

You can coordinate autonomous AI agents across multiple repositories using multi-repo orchestration with isolated work trees and safe merges. This automates planning, coding, reviewing, and verification from a natural-language goal.

What is runtime plan mutation in autonomous workflow orchestration?

Runtime plan mutation allows an autonomous AI engineering workflow to dynamically adapt its execution DAG during processing. It enables graceful degradation with explicit debt tracking when resolving software engineering tasks.

How do I trigger an end-to-end software build from a natural-language goal?

To trigger an end-to-end software build, post a plan request to the control plane with your natural-language goal and repository details. This single API call initiates planning, coding, merging, and verification.

Do I need durable checkpoints for autonomous AI engineering workflows?

Yes, durable checkpoints are required for resilient execution in autonomous AI engineering workflows. They ensure durable state recovery alongside a governance layer and configurable runtime to orchestrate software delivery.

Can I use level-based execution for parallel workstreams in a single repository?

Yes, level-based execution supports parallel workstreams in both single-repo and multi-repo environments. This workflow orchestration manages isolated work trees to safely coordinate autonomous AI agents.

What are the limitations of automating software engineering with AI agents?

While autonomous AI agents automate planning and coding, complex multi-repo merges may require explicit debt tracking and graceful degradation. Runtime plan mutation helps manage failures, but a governance layer remains essential for verification.