harnesscp-context

Restructure Copilot project instructions into layered always-on and conditional files.

Updated Mar 31, 2026
One-click install
npx skills add https://github.com/LSDIPPOLLC/harness-cp --skill harnesscp-context
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: harnesscp-context
Source: https://github.com/LSDIPPOLLC/harness-cp/tree/main/.github/skills/harnesscp-context
Command: npx skills add https://github.com/LSDIPPOLLC/harness-cp --skill harnesscp-context

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill helps you fix overloaded, scattered, or ineffective Copilot context by deciding what should stay always-on, what should load conditionally, and what should move into skills, memory, or references.

Core Features & Use Cases

  • Context Budget Analysis: Assess whether project instructions are too sparse, balanced, or bloated for reliable model behavior.
  • Instruction Restructuring: Split broad guidance into focused layers so the model sees only what matters for the current task.
  • Progressive Disclosure Design: Plan conditional instruction files for specific file types, domains, or workflows.
  • Use Case: If a repository’s Copilot instructions are too long and the model keeps missing key rules, this Skill helps reorganize the guidance so critical information stays visible while domain details load only when needed.

Quick Start

Use the harnesscp-context skill to review the repository’s Copilot instructions and reorganize them into a lean always-on core plus targeted conditional files.

Frequently Asked Questions about harnesscp-context

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

FAQPage Schema
How do I reduce wasted tokens in overloaded Copilot instructions?

To reduce wasted tokens in overloaded Copilot instructions, restructure project guidance by splitting broad rules into focused layers. This ensures the model sees only what matters for the current task while critical information stays visible.

What is progressive disclosure design for AI context engineering?

Progressive disclosure design in context engineering plans conditional instruction files for specific file types, domains, or workflows. This approach loads domain details only when needed, keeping the always-on core lean and improving model focus.

How do I organize a repository's Copilot instructions into always-on and conditional files?

To organize Copilot instructions, perform a context inventory to deduplicate and validate guidance. Decide what belongs in always-on files, conditional files, skills, memory, or references to preserve critical rules while reducing bloat.

Why does Copilot keep missing key rules when my project instructions are too long?

Copilot misses key rules when project instructions are too long because the context budget is bloated, diluting model focus. Restructuring guidance into targeted conditional layers ensures critical information stays visible instead of being ignored.

When do I need a context budget analysis for Copilot instruction restructuring?

You need a context budget analysis when assessing whether project instructions are too sparse, balanced, or bloated for reliable model behavior. It evaluates if your current guidance volume causes the model to lose focus on critical rules.

Can I use context engineering to split broad Copilot guidance into focused layers?

Yes, context engineering splits broad Copilot guidance into focused instruction layers. By applying progressive disclosure design, you separate always-on core rules from conditional domain details to optimize token usage and model reliability.