.github-private

This is a special repository for the Universal-Standard organization that enables a powerful, lesser-known GitHub feature: Custom AI Agents.

The .github-private repository has a unique capability to host custom AI agent profiles that extend GitHub Copilot's functionality with organization-specific expertise and workflows.

Custom AI agents are specialized AI assistants that can be configured to understand your team's specific workflows, coding standards, and requirements. They appear as Markdown files with YAML frontmatter in the .github/agents directory.

• 🔒 Private & Secure: Agents in .github-private are only accessible to organization members
• 🧪 Safe Testing: Test and iterate on agents privately before releasing org-wide
• 🎯 Specialized Knowledge: Create agents with domain-specific expertise
• 🔧 Tool Integration: Agents can access shell commands and external tools
• 📦 Easy Deployment: Move tested agents to organization-level when ready

.github-private/
├── README.md
└── .github/
    └── agents/
        ├── code-reviewer.md
        ├── documentation-specialist.md
        ├── kubernetes-specialist.md
        ├── python-expert.md
        ├── testing-expert.md
        └── [specialized ecosystem agents...]

File: .github/agents/documentation-specialist.md

An expert in creating and improving README files and documentation. Use this agent when you need help with:

• Writing comprehensive README files
• Creating API documentation
• Improving documentation clarity
• Adding usage examples

File: .github/agents/code-reviewer.md

A specialized agent for conducting thorough code reviews with focus on:

• Code quality and maintainability
• Security vulnerabilities
• Performance optimization
• Testing coverage

File: .github/agents/testing-expert.md

Focused on software testing and test-driven development:

• Writing unit and integration tests
• Ensuring test quality
• Following testing best practices
• Multiple language/framework support

File: .github/agents/kubernetes-specialist.md

Expert in Kubernetes and container orchestration:

• Creating and validating K8s manifests
• Helm chart development
• Security best practices
• Resource optimization

The repository also includes specialized agents under .github/agents/ for the broader Universal Standard ecosystem:

• typescript-architect.md — TypeScript architecture and shared patterns
• api-integration-specialist.md — API contracts and service integrations
• security-hardening-specialist.md — secure defaults, auth, and secrets hygiene
• workflow-orchestrator.md — multi-step and multi-agent workflow coordination
• architecture-architect.md — cross-system architecture and boundaries
• python-platform-specialist.md — Python platform and service design
• frontend-ux-specialist.md — user-facing flows and frontend quality
• devops-release-engineer.md — build, deploy, and release operations
• data-model-specialist.md — domain models and shared data contracts
• debugging-root-cause-specialist.md — regression and incident root-cause analysis
• ci-cd-specialist.md — pipeline automation and delivery reliability
• database-specialist.md — schema, migrations, and query strategy
• observability-specialist.md — logs, metrics, traces, and alerting
• prompt-and-policy-specialist.md — prompt design, guardrails, and policy-aware agents
• interop-specialist.md — compatibility across protocols, repos, and shared interfaces
• monorepo-maintainer.md — large-repo structure and developer workflows
• legacy-migration-specialist.md — staged modernization and upgrade planning
• quality-assurance-lead.md — cross-project validation strategy and release confidence
• infra-platform-specialist.md — shared infrastructure and platform reliability
• release-notes-specialist.md — changelogs, rollout notes, and release summaries

Agents in this .github-private repository are automatically available to organization members for testing:

1. Via GitHub Copilot Chat: Select the agent from the agent picker
2. Via Copilot CLI: Use gh copilot with agent selection
3. In PRs and Issues: Reference agents in conversations

Once an agent is tested and validated:

1. Move the agent profile from .github-private/.github/agents/ to the organization's main agents directory
2. This makes the agent available across all repositories in the organization
3. All organization members can then access the agent

Each agent is a Markdown file with YAML frontmatter:

---
name: Agent Name
description: Brief description of the agent's purpose
tools:
  - shell
  - other-tool
---

# Agent Name

You are a specialized AI agent focused on [specific domain].

## Core Responsibilities

- List key responsibilities
- Define scope of expertise

## Guidelines

1. Specific rules and principles
2. Best practices to follow
3. Quality standards

## Focus Areas

- Specific topics or tasks
- Technologies or frameworks

[Additional instructions and context...]

1. Clear Scope: Define what the agent should and shouldn't do
2. Specific Instructions: Provide detailed guidelines and examples
3. Consistency: Maintain consistent tone and approach
4. Tools Access: Only grant necessary tool permissions
5. Iterative Development: Start simple, refine based on usage

Create a "Developer Onboarding" agent that knows your team's setup process, coding standards, and common gotchas.

Build an "Architecture Advisor" agent familiar with your system design patterns and technology stack.

Develop a "Security Auditor" agent that checks code against your organization's security policies.

Create agents for specific frameworks your team uses (React, Spring Boot, etc.) with your team's conventions.

• GitHub Docs: Testing and releasing custom agents
• GitHub Docs: Creating custom agents
• GitHub Blog: Copilot CLI

1. Review the example agents in .github/agents/
2. Test these agents with your workflows
3. Create custom agents specific to your team's needs
4. Share feedback and iterate on agent designs
5. Release successful agents to the organization

Note: This feature is part of GitHub Copilot for Business/Enterprise and requires appropriate licensing and permissions.