Skip to content

README.md

Latest commit

 

History

History
107 lines (81 loc) · 5.72 KB

README.md

File metadata and controls

107 lines (81 loc) · 5.72 KB

🤖 GitHub Copilot Custom Agents

Specialized agent configurations for CIA Compliance Manager (v1.1.43).

Context & Setup

All agents read these files first:

  1. README.md — Project context, features, architecture
  2. .github/workflows/copilot-setup-steps.yml — Node 25, npm, build/test commands
  3. .github/copilot-mcp.json — MCP servers (filesystem, github, memory, sequential-thinking, playwright)

Tech Stack

TypeScript 6.0.2 · React 19.x · Vite 8 · Vitest 4.x · Cypress 15.x · ESLint 10.x · TailwindCSS 4 · Node ≥25 · ES2025

Skills Framework

Strategic skills in .github/skills/ guide all agents:

Priority Skill File
MANDATORY 🔐 Security by Design security-by-design.md
CRITICAL ✨ Code Quality Excellence code-quality-excellence.md
MANDATORY 🛡️ ISMS Compliance isms-compliance.md
MANDATORY 🧪 Testing Excellence testing-excellence.md
IMPORTANT ⚡ Performance Optimization performance-optimization.md
IMPORTANT 🎨 UI/UX Design System ui-ux-design-system.md
IMPORTANT ♿ Accessibility Excellence accessibility-excellence.md
IMPORTANT 📝 Documentation Standards documentation-standards.md
ADVISORY 🎯 Product Quality Analysis product-quality-analysis.md
MANDATORY 🏗️ C4 Architecture c4-architecture-documentation.md
ADVISORY 🤖 Agentic Workflows github-agentic-workflows.md
ADVISORY 🔌 MCP Integration mcp-server-integration.md

Skills vs Agents: Skills define what to do (principles). Agents define how to do it (execution).

Available Agents

Agent Purpose When to Use
@code-review-agent Code quality, security, reusability PR reviews, quality checks
@documentation-agent Technical docs, API docs, diagrams Documentation updates, JSDoc
@product-task-agent Issue creation, quality coordination Product analysis, issue triage
@security-compliance-agent Security review, ISMS compliance Security audits, threat models
@testing-agent Vitest/Cypress testing, coverage Writing and improving tests
@typescript-react-agent TypeScript/React development Components, hooks, types

Agent Selection Guide

Need code changes?
├── TypeScript/React → @typescript-react-agent
├── Tests → @testing-agent
└── Security fix → @security-compliance-agent

Need review?
├── Code quality → @code-review-agent
├── Security audit → @security-compliance-agent
└── Documentation → @documentation-agent

Need coordination?
└── Issues/planning → @product-task-agent

Key Rules (All Agents)

  1. MUST check existing code before creating new (reusability is CRITICAL)
  2. MUST use explicit TypeScript types (no any)
  3. MUST maintain 80%+ test coverage
  4. MUST follow ISMS compliance requirements
  5. MUST run npm run lint && npm run test && npm run build before committing

Contributing

When creating or modifying agents:

  • Follow GitHub custom agents documentation
  • Include YAML frontmatter (name, description, tools: ["*"])
  • Repo-level agents MUST NOT include mcp-servers blocks — MCP servers are configured centrally in .github/copilot-mcp.json
  • Reference relevant skills from .github/skills/
  • Keep agents focused and concise — aim below ~200 lines per agent

Copilot Coding Agent Workflow

Repo-level agents coordinate with Copilot coding agent using these MCP tools (configured in .github/copilot-mcp.json):

Tool Purpose Key Parameters
assign_copilot_to_issue Assign issue to Copilot coding agent base_ref, custom_instructions
create_pull_request_with_copilot Create PR with Copilot implementing changes base_ref, problem_statement
get_copilot_job_status Track progress id (PR number / job id)

Stacked PRs: use base_ref: "copilot/issue-<NNN>" to build incremental changes on a prior Copilot PR branch.

Policy → Agent Matrix

ISMS Policy Primary Agent Supporting Agent
Information Security Policy @security-compliance-agent @code-review-agent
Secure Development Policy @security-compliance-agent @typescript-react-agent, @testing-agent, @code-review-agent
Open Source Policy @code-review-agent @security-compliance-agent
Vulnerability Management @security-compliance-agent @testing-agent
Data Classification @security-compliance-agent @typescript-react-agent
AI Policy / OWASP LLM @security-compliance-agent @product-task-agent
Change Management @product-task-agent @documentation-agent

Made with ❤️ for CIA Compliance Manager | Hack23 AB