CLAUDE.md TemplatesCLAUDE.md Template

Angular Material + FastAPI Multi-Agent App CLAUDE.md Template

CLAUDE.md Template for Angular Material + FastAPI multi-agent app starter with a copyable Claude Code block for Claude Code.

CLAUDE.md TemplateAngular MaterialFastAPIMulti-Agent AppClaude CodeFrontendBackendDockerKubernetesOpenAPIJWT

Target User

Frontend and Backend developers building an Angular Material + FastAPI multi-agent app

Use Cases

  • Generate a copyable CLAUDE.md template for an Angular Material + FastAPI multi-agent project
  • Document architecture decisions for multi-agent orchestration
  • Provide a ready-to-paste Claude Code block for project setup and rules

Markdown Template

Angular Material + FastAPI Multi-Agent App CLAUDE.md Template

# CLAUDE.md

Project role
- You are Claude Code, a helper for Angular Material + FastAPI multi-agent app starters. You produce precise rules, directory structures, and code templates for frontend and backend agents and orchestrators.

Architecture rules
- Use a mono-repo with a distinct frontend and backend folder, plus an agents folder for orchestration logic.
- Frontend communicates with backend via RESTful OpenAPI endpoints; backend exposes a single gateway for agent coordination.
- Agents run as isolated components with well-defined input/output contracts.
- Separate concerns: UI, API, and agent orchestration must not be tightly coupled.

File structure rules
- Do not create irrelevant folders; include only frontend, backend, and agents under a single repo. Example structure:
  frontend/
    src/
      app/
        components/
        services/
      index.html
      main.ts
    angular.json
  backend/
    app/
      main.py
      api/
      models/
      schemas/
      services/
      agents/
    requirements.txt
    Dockerfile
  agents/
    orchestrator/
    agent-a/
  infra/
    docker-compose.yml

Authentication rules
- Use OAuth2 Password or Authorization Code with PKCE for the frontend; FastAPI should implement OAuth2 and JWT tokens.
- Never embed plaintext secrets in code or CLAUDE.md template outputs. Use environment variables in deployment.

Database rules
- Use PostgreSQL; use SQLAlchemy or FastAPI-compatible ORM; apply Alembic migrations.
- Do not store secrets in the database; encrypt sensitive fields if needed.

Validation rules
- Use Pydantic models in FastAPI for request/response validation.
- Validate all inputs at the boundary; fail fast with meaningful errors.

Security rules
- Enable CORS for the frontend origin only.
- Enforce TLS in production; do not serve over plain HTTP.
- Validate tokens on every protected endpoint; implement proper auth scopes for agents.

Testing rules
- Backend: unit tests for services and models; API integration tests with the TestClient.
- Frontend: unit tests for components/services; end-to-end tests with Cypress if used.
- Run tests in CI on push to main or PRs.

Deployment rules
- Use Docker Compose for local dev; include services for frontend, backend, and a Postgres container.
- For production, prepare Kubernetes manifests or a robust deployment pipeline.
- Do not deploy without secret management and proper config maps.

Things Claude must not do
- Do not hardcode secrets or credentials.
- Do not bypass authentication checks.
- Do not introduce unapproved dependencies or libraries.
- Do not assume network access to internal services beyond the defined architecture.

Overview

Direct answer: This CLAUDE.md template provides a ready-to-paste Claude Code starter for an Angular Material frontend and FastAPI backend that coordinate multiple agents. It enforces a stack-specific structure, security, and testing practices so developers can bootstrap a scalable multi-agent app with Claude Code guidance.

This page is a copyable CLAUDE.md template page for the Angular Material + FastAPI stack. It supplies concrete constraints, file layout, and Claude Code rules that translate directly into starter projects and starter documentation.

When to Use This CLAUDE.md Template

  • Starting a new Angular Material UI paired with a FastAPI backend that orchestrates multiple AI agents.
  • Documenting architecture and coding conventions for a multi-agent workflow.
  • Providing a paste-ready Claude Code block to enforce project structure and rules across teams.
  • Ensuring security, validation, and testing practices are baked into the starter scaffold.

Copyable CLAUDE.md Template

# CLAUDE.md

Project role
- You are Claude Code, a helper for Angular Material + FastAPI multi-agent app starters. You produce precise rules, directory structures, and code templates for frontend and backend agents and orchestrators.

Architecture rules
- Use a mono-repo with a distinct frontend and backend folder, plus an agents folder for orchestration logic.
- Frontend communicates with backend via RESTful OpenAPI endpoints; backend exposes a single gateway for agent coordination.
- Agents run as isolated components with well-defined input/output contracts.
- Separate concerns: UI, API, and agent orchestration must not be tightly coupled.

File structure rules
- Do not create irrelevant folders; include only frontend, backend, and agents under a single repo. Example structure:
  frontend/
    src/
      app/
        components/
        services/
      index.html
      main.ts
    angular.json
  backend/
    app/
      main.py
      api/
      models/
      schemas/
      services/
      agents/
    requirements.txt
    Dockerfile
  agents/
    orchestrator/
    agent-a/
  infra/
    docker-compose.yml

Authentication rules
- Use OAuth2 Password or Authorization Code with PKCE for the frontend; FastAPI should implement OAuth2 and JWT tokens.
- Never embed plaintext secrets in code or CLAUDE.md template outputs. Use environment variables in deployment.

Database rules
- Use PostgreSQL; use SQLAlchemy or FastAPI-compatible ORM; apply Alembic migrations.
- Do not store secrets in the database; encrypt sensitive fields if needed.

Validation rules
- Use Pydantic models in FastAPI for request/response validation.
- Validate all inputs at the boundary; fail fast with meaningful errors.

Security rules
- Enable CORS for the frontend origin only.
- Enforce TLS in production; do not serve over plain HTTP.
- Validate tokens on every protected endpoint; implement proper auth scopes for agents.

Testing rules
- Backend: unit tests for services and models; API integration tests with the TestClient.
- Frontend: unit tests for components/services; end-to-end tests with Cypress if used.
- Run tests in CI on push to main or PRs.

Deployment rules
- Use Docker Compose for local dev; include services for frontend, backend, and a Postgres container.
- For production, prepare Kubernetes manifests or a robust deployment pipeline.
- Do not deploy without secret management and proper config maps.

Things Claude must not do
- Do not hardcode secrets or credentials.
- Do not bypass authentication checks.
- Do not introduce unapproved dependencies or libraries.
- Do not assume network access to internal services beyond the defined architecture.

Recommended Project Structure

frontend/
  angular.json
  package.json
  src/
    main.ts
    index.html
    app/
      app.module.ts
      components/
      services/
backend/
  app/
    main.py
    api/
      v1/
    models/
      __init__.py
    schemas/
    services/
  requirements.txt
  Dockerfile
agents/
  orchestrator/
    __init__.py
    orchestrator.py
  agent-a/
    __init__.py
    agent.py
infra/
  docker-compose.yml
  k8s/
    deployment.yaml
README.md

Core Engineering Principles

  • Clear separation of frontend UI, backend API, and agent orchestration logic
  • Single source of truth for data shapes and access patterns
  • Idempotent operations and deterministic agent coordination
  • Explicit contracts via OpenAPI for backend and inter-agent communication
  • Test-driven stance with guardrails for architecture drift

Code Construction Rules

  • Frontend TypeScript should use Angular Material components with well-typed services
  • Backend FastAPI endpoints must include explicit Pydantic models for inputs and outputs
  • Agent orchestration should expose stable interfaces; avoid tight coupling to specific agents
  • Environment variables used for config; never rely on hardcoded values
  • Code should be modular; avoid large monoliths; favor small composable services
  • All data transfer objects must be validated at entry points
  • Use OpenAPI generated client/model stubs for inter-service communication
  • Do not rely on non-approved libraries or insecure dependencies

Security and Production Rules

  • Enable TLS, use secure cookies, and restrict CORS origins to the frontend domain
  • Implement robust auth with JWTs and token refresh; validate scopes for each endpoint
  • Encrypt or hash sensitive database fields where appropriate
  • Limit request rates per endpoint to prevent abuse; monitor with simple logs and metrics
  • Do not log sensitive payloads; redact secrets in logs

Testing Checklist

  • Unit tests for backend services and Pydantic models
  • API integration tests covering agent interaction flows
  • Frontend component and service tests; ensure Angular Material components render correctly
  • End-to-end tests for agent orchestration path
  • CI checks for lint, type checks, and security scanners

Common Mistakes to Avoid

  • Mixing frontend and backend logic in the same module; keep concerns separated
  • Hardcoding secrets or credentials in code or CLAUDE.md outputs
  • Over-privileging agents or endpoints; apply least privilege
  • Relying on deprecated dependencies or insecure configurations

FAQ

What is this CLAUDE.md Template used for
A copyable starter for an Angular Material + FastAPI multi-agent app that Claude Code can populate and enforce.
Can I adapt this template to other stacks
Yes, but update the architecture rules, file structure, and dependencies to match the new stack.
How do I run the starter locally
Follow the recommended project structure, install dependencies in frontend and backend, start backend, then start frontend.
What security practices are required
Use TLS, OAuth2/JWT, restricted CORS, token validation, and avoid embedding secrets in code.
What should Claude avoid in this template
Do not introduce unapproved libraries, hardcode secrets, or bypass authentication and input validation.
---

Notes: This page uses the exact phrase CLAUDE.md template in the copyable block and throughout the page to align with search expectations for CLAUDE.md Template pages.