swe-workflow

Automate triage-first planning and step-by-step execution for multi-file code tasks.

Updated Mar 20, 2026
One-click install
npx skills add https://github.com/ex-git/swe-workflow --skill swe-workflow
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: swe-workflow
Source: https://github.com/ex-git/swe-workflow/tree/main
Command: npx skills add https://github.com/ex-git/swe-workflow --skill swe-workflow

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

The SWE workflow skill provides a rigorous process to ensure code tasks are triaged, planned, executed, verified, and reflected upon, preventing scope creep and poor quality in AI-assisted coding tasks.

Core Features & Use Cases

  • Enforces mandatory triage before edits
  • Structures task planning with per-step checks and evidence
  • Supports full workflow for broad changes and multi-file tasks
  • Maintains plan, context, and optional repo-memory for handoffs

Quick Start

Declare workflow mode (Lightweight or Full) and generate a plan using the canonical template.

Frequently Asked Questions about swe-workflow

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

FAQPage Schema
How do I enforce a triage-first workflow for AI coding agents?

The skill automates a triage-first workflow for AI coding agents by blocking edits until a formal triage phase evaluates and scopes the task, ensuring structured execution.

What is the best way to manage multi-file code tasks with AI?

The best way to manage multi-file code tasks with AI is using a structured workflow that generates a canonical plan and maintains context-aware handoffs, preventing scope creep across complex edits.

How do I create a structured plan for AI-assisted code changes?

You create a structured plan for AI-assisted code changes by applying a canonical template that breaks broad tasks into step-by-step execution with per-step checks and evidence.

Does AI-assisted coding workflow support context handoffs between sessions?

Yes, AI-assisted coding workflow supports context handoffs between sessions by maintaining plans, context, and optional repo-memory to preserve state across multi-file task execution.

Can I use a lightweight planning mode for smaller AI coding tasks?

Yes, you can declare a lightweight planning mode for smaller AI coding tasks, offering a less rigid workflow compared to the full mode required for broad, multi-file changes.

Why does AI coding scope creep happen and how to prevent it?

AI coding scope creep happens due to unmanaged edits and poor planning, and you prevent it by enforcing a rigorous process of triage, step-by-step verification, and reflection upon task completion.