skills.rest
microsoft
microsoftOFFICIAL
Explore all skills by this author →
Website𝕏@OpenAtMicrosoftGitHub

starting-nori

Establishes the Nori workflow, branch protection checks, operating modes, and coding guidelines before any task.

4.9k|585|Updated Aug 22, 2019
Search Tags:#Test-Driven Development#coding guidelines#Planning#Workflow#branch protection#agent modes#Git worktree
One-click install
npx skills add https://github.com/microsoft/FluidFramework --skill starting-nori
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: starting-nori
Source: https://github.com/microsoft/FluidFramework/tree/main/.agency/plugins/nori/skills/starting-nori
Command: npx skills add https://github.com/microsoft/FluidFramework --skill starting-nori

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

It prevents unstructured AI coding sessions by enforcing a consistent startup workflow: checking for protected branches, selecting an operating mode, and loading coding guidelines before any work begins.

Core Features & Use Cases

  • Protected Branch Check: Detects when you are on main, master, or dev and prompts you to create a branch or git worktree before making changes.
  • Dual Operating Modes: Offers nori-full-send mode for autonomous plan-then-execute work, or nori-copilot mode where each step requires your explicit approval.
  • Opinionated Coding Guidelines: Enforces YAGNI, test-driven development, root-cause debugging, and a direct non-deferential communication tone.
  • Use Case: Start any coding session in a repository and the agent will verify your branch, ask your preferred mode, research with subagents, write a plan, and follow TDD before touching code.

Quick Start

Ask the agent to start working on any coding task and it will load the Nori workflow, check your git branch, and prompt you to choose full-send or copilot mode.

Frequently Asked Questions about starting-nori

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

FAQPage Schema
How do I enforce a consistent workflow for AI coding agents?▼

Load a startup skill that runs before any task, checking git branch state, asking the user to pick an operating mode, and adding required steps like planning and TDD to a todo list. This skill does exactly that for every user query.

What is the difference between nori-full-send and nori-copilot mode?▼

In nori-full-send mode the agent creates a plan with you, then works autonomously through TDD, docs, and final checks. In nori-copilot mode it telegraphs each step and asks permission before continuing, including optional TDD, docs, and PR creation.

How do I stop an AI agent from committing directly to main?▼

This skill checks git status at session start and, if you are on main, master, dev, or a similar protected branch, asks whether to create a new branch or a git worktree before any changes are made.

Does this workflow support test-driven development?▼

Yes. Full-send mode mandates reading and following a test-driven-development skill, writing tests before implementation. Copilot mode asks the user whether to apply TDD before proceeding.

When should a subagent skip this startup skill?▼

Subagents dispatched to execute a specific task should skip this skill entirely, as indicated by the SUBAGENT-STOP block. It is designed only for top-level interactive sessions, not delegated task execution.

Related Skills

Explore More Software Engineering Skills→
open-haxopen-hax
6

git-safety-check

Check git branch, working tree, remotes, and submodules before destructive operations.

Official
Intermediate
coco-researchcoco-research
264

finishing-a-development-branch

Verify tests, then merge, pull request, keep, or discard a development branch.

Community
Advanced
dmulcaheydmulcahey

systematic-debugging

Trace software bugs to root causes using a four-phase investigation methodology.

Community
Advanced
plurigridplurigrid
60

babashka

Execute lightweight Clojure scripts with Babashka without JVM startup.

Official
Basic