Skip to main content

In this article

Customizing HVE Core

Before You Customize

Your installation method determines which customization options are available.

Customization LevelExtension OnlyInstaller Skill (Clone)Direct Clone
VS Code SettingsYesYesYes
copilot-instructions.mdYesYesYes
.instructions.md filesYes (your repo)YesYes
Multi-kind selectionNoYesYes
Modify agentsNoNoYes
Modify prompts and skillsNoNoYes
Build system changesNoNoYes
Fork and extendNoNoYes

The HVE Core extension installs the complete active component set. For MCP guidance, installation-method selection, or selective cloning, ask an agent to use the included hve-core-installer skill. Its complete and custom flows select agents, prompts, instructions, and distributable skill directories while preserving repository-relative paths. Hooks are not copied. For full artifact modification, use a clone-based installation method.

Customization Spectrum

HVE Core supports a range of customization depths. Start with the lightest option that meets your needs, then move deeper when the situation demands it.

graph LR
A["VS Code Settings"] --> B["Instructions"]
B --> C["Agents & Prompts"]
C --> D["Skills"]
D --> E["Plugin Manifest"]
E --> F["Build System"]
F --> G["Fork & Extend"]

style A fill:#e8f5e9
style B fill:#c8e6c9
style C fill:#a5d6a7
style D fill:#81c784
style E fill:#66bb6a
style F fill:#4caf50
style G fill:#388e3c
ApproachDescription
VS Code SettingsIndividual preferences like font size, theme, and editor behavior. No files to create or share.
InstructionsConfigure Copilot behavior through .github/copilot-instructions.md and .instructions.md files. Lowest effort with highest return for shaping AI output.
Agents and PromptsSpecialized workflows: agents for multi-turn interactions, prompts for single-shot tasks. Both accept tool restrictions and delegation rules.
SkillsDomain knowledge in self-contained bundles with optional scripts. Use when instruction files alone cannot capture the depth of a domain.
Plugin ManifestSynchronize the complete plugin and VSIX component set from tracked source.
Build SystemValidation scripts, schema checks, and extension packaging.
Fork and ExtendFull control over every artifact. Fork the repository when your changes diverge significantly from upstream.

Choose Your Approach

GoalApproachFiles InvolvedDifficulty
Set coding standards for CopilotInstructions.github/copilot-instructions.md, .instructions.mdLow
Create a reusable workflowPrompt.github/prompts/{package-id}/name.prompt.mdLow
Build a specialized Copilot assistantAgent.github/agents/{package-id}/name.agent.mdMedium
Package domain expertiseSkill.github/skills/{package-id}/{skill}/SKILL.mdMedium
Change managed distribution membershipPlugin Manifest.github/plugin.json, docs/plugins/hve-core.mdMedium
Add custom validation or packagingBuild Systemscripts/, package.jsonHigh
Diverge from upstream entirelyFork and ExtendFull repositoryHigh

Authoring with HVE Builder

Use the hve-builder skill to create, improve, refactor, replace, review, or validate prompts, instructions, agents, subagents, and skills. It resolves the write boundary, runs independent static review, and applies one behavior gate with route-specific execution. Major mutations and behavior-bearing review targets execute testing, while eligible no-runtime review targets and Minor or Medium mutations are satisfied-and-skipped. Known target files and caller-supplied canonical references remain bounded lifecycle reads; open-ended exploration and decision-critical research activate rpi-research.

The retained prompt-builder, prompt-analyze, and prompt-refactor skills remain compatibility aliases for legacy requests. They route to hve-builder and do not own separate authoring workflows.

Each artifact guide below includes an "Authoring with HVE Builder" section with type-specific examples.

Role-Based Entry Points

Each HVE role benefits from different customization techniques. The table below maps the nine roles to the guides most relevant to their workflow.

RoleRecommended GuidesRationale
EngineerInstructions, AgentsCoding standards and specialized review agents accelerate daily development
TPMPrompts, Plugin ManifestReusable planning prompts and managed distribution standardize project workflows
Tech Lead / ArchitectInstructions, Agents, SkillsStandards enforcement, architecture review agents, and deep domain knowledge
Security ArchitectSkills, InstructionsCompliance knowledge packages and security-focused coding conventions
Data ScientistSkills, PromptsAnalytical domain bundles and repeatable notebook workflows
SRE / OperationsInstructions, Environment, Local TelemetryInfrastructure conventions, DevContainer tuning, and local telemetry workflows
Platform / ObservabilityCopilot OTel Metrics, Local TelemetryAgent usage, token cost, and latency measurement through OpenTelemetry
Business Program ManagerPrompts, Team AdoptionSprint-planning prompts and governance patterns for stakeholder alignment
New ContributorInstructions, EnvironmentQuick onboarding through conventions and a ready-to-use development environment
UtilityPlugin Manifest, Build SystemCross-cutting tooling assembly and validation pipeline customization

File Index

  1. Customizing with Instructions: Configure Copilot with copilot-instructions.md and instruction files
  2. Creating Custom Agents: Build specialized agents with tool restrictions and subagent delegation
  3. Creating Custom Prompts: Author reusable prompt templates with variables
  4. Authoring Custom Skills: Create domain knowledge packages
  5. Managing the HVE Core Plugin Manifest: Synchronize one plugin and VSIX membership
  6. Build System and Validation: Manifest sync, schema validation, npm scripts
  7. Forking and Extending: Full fork-and-extend customization
  8. Environment Customization: DevContainers, VS Code settings, MCP servers
  9. Team Adoption and Governance: Governance, naming, onboarding, change management
  10. Local Telemetry: Enable local telemetry, review capture and storage schema mechanics, and generate reports
  11. Copilot OpenTelemetry Metrics: Export Copilot OTel signals to a local Grafana stack, or to a fleet-wide Azure pipeline, and query agent, token, and latency data

🤖 Crafted with precision by ✨Copilot following brilliant human instruction, then carefully refined by our team of discerning human reviewers.