refine

Align requirements, glossary terms, and interface contracts via a codebase-first grill-me protocol.

Updated Mar 2, 2022
One-click install
npx skills add https://github.com/ysuurme/azure_hello_world --skill refine-ysuurme
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: refine
Source: https://github.com/ysuurme/azure_hello_world/tree/main/.agents/skills/refine
Command: npx skills add https://github.com/ysuurme/azure_hello_world --skill refine-ysuurme

SYSTEM DOCUMENTATION & REQUIREMENTS

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

What problem does it solve?

Refine prevents wasted engineering cycles caused by unclear shared language, unagreed boundaries, and missing interface contracts by forcing early, codebase-first alignment before any implementation begins.

Core Features & Use Cases

  • Grill-me protocol for alignment: Runs a full decision-tree that resolves glossary terms, architectural boundaries, constraints, and out-of-scope decisions prior to building.
  • Context-first system contracts: Produces a structured Agent Brief with interface-first schemas, error states, and testability notes rather than implementation plans.
  • Architecture decision capture: Converts every architectural decision into MADR-format ADRs and updates contextual invariants and constraints.
  • Session discipline and guardrails: Uses context sensitivity thresholds to avoid drifting alignment when the context window is large or unstable.

Quick Start

Ask the AI to run the refine skill to produce an updated CONTEXT.md, an Agent Brief with interface contracts, and any required MADR ADRs for your feature request.

Frequently Asked Questions about refine

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

FAQPage Schema
How do I align requirements and interface contracts before coding?

To align interface contracts before coding, run a grill-me protocol that resolves glossary terms and architectural boundaries, producing an interface-first Agent Brief and MADR ADRs.

What is the best way to maintain architectural alignment during a refactor?

Maintaining architectural alignment during a refactor involves capturing every architectural decision in MADR-format ADRs and updating contextual invariants in your CONTEXT.md file to preserve bounded contexts.

When do I need to define bounded contexts and glossary terms for a new service?

You need to define bounded contexts and glossary terms for a new service when shared language is ambiguous or missing and architectural constraints must be preserved prior to implementation.

How do I prevent context window drift in an agentic workflow?

To prevent context window drift in an agentic workflow, apply context sensitivity thresholds as guardrails to avoid drifting alignment when the context window becomes large or unstable.

Does this alignment protocol require existing documentation to work?

Yes, this alignment protocol requires reading the existing codebase and existing CONTEXT.md or AGENTS.md files to produce an updated CONTEXT.md and structured interface contracts.