next

Identifies the optimal next step from current Git repository state and thread context.

13|3|Updated Mar 2, 2026
One-click install
npx skills add https://github.com/misty-step/harness-kit --skill next-misty-step
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: next
Source: https://github.com/misty-step/harness-kit/tree/main/skills/next
Command: npx skills add https://github.com/misty-step/harness-kit --skill next-misty-step

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Identify the most actionable next-step given the current thread and repository state.

Core Features & Use Cases

  • Grounded recommendations: provide a single best next move based on live context, separating agent-owned actions from user-owned actions.
  • Scope-aware planning: handles repo status, branches, backlog, and open tasks to support scheduling and execution.
  • Use Case: during a development cycle, ask for /next to decide whether to groom backlog, ship a feature, or verify status before proceeding.

Quick Start

Ask me for the best next move based on the current thread and repository state.

Frequently Asked Questions about next

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

FAQPage Schema
How do I decide the best next move during a live coding session?

To decide the best next move during a live coding session, this skill analyzes current thread and repository state to output a single actionable step, separating agent-owned actions from user-owned actions.

What is the best way to prioritize backlog grooming versus feature delivery?

The best way to prioritize backlog grooming versus feature delivery is by evaluating open tasks and repo status, which generates a grounded recommendation on whether to schedule execution or verify status first.

Can I use branch and tracker status for scope-aware planning?

Yes, you can use branch and tracker status for scope-aware planning, as this skill handles repository status and open tasks across various branches to support scheduling and execution decisions.

How does decision-making work when separating agent-owned and user-owned actions?

Decision-making works by applying grounding rules to the current state and enforcing a clear separation between agent-owned and user-owned actions, ensuring each output is a single best move with defined stop conditions.

When should I not use automated next-step recommendations for repo management?

You should not use automated next-step recommendations for repo management if you lack a live thread context or active repository state, as the output depends entirely on grounded, real-time data to function correctly.