A Claude Code plugin that brings a full multi-agent development team and an extensive knowledge base for the Water Framework directly into your IDE.
The plugin ships 5 specialized agents and 11 skills that cover every phase of Water Framework development: requirements analysis, solution architecture, code scaffolding, backend implementation, and test generation — all without needing to read the framework source code.
Requires Claude Code with
/pluginsupport.
# 1. Add this repo as a marketplace
/plugin marketplace add Water-Framework/claude
# 2. Install the plugin
/plugin install water-framework@Water-Framework/claudeAlternatively, add it as a git submodule inside an existing project:
git submodule add https://github.com/Water-Framework/claude.git .claudeAgents are autonomous specialists invoked automatically by Claude Code based on the task at hand, or called directly by name.
Functional analysis — Decomposes complex or ambiguous requirements into structured, actionable specifications aligned with Water Framework architectural principles. Invoke before any code is written when requirements span multiple bounded contexts or involve cross-aggregate state.
Technical architecture — Produces Technical Architecture Analyses (TAA): which modules/microservices are needed, how they communicate, which runtime each targets, and integration blueprints. Does not write code — it defines what gets built and how the pieces connect.
Code generation — Executes all yo water:* generator commands: new projects, add entities, add REST services, build, publish. Enforces the generator-first protocol — no manual file creation where the generator can do it.
Backend implementation — Implements code inside existing generated module structures: entity fields, business logic, custom methods, security annotations, custom queries, and custom REST endpoints. Works only within what the generator has already produced.
Test generation — Writes JUnit 5 unit tests and Karate REST integration tests targeting ≥80% instruction coverage for SonarQube compliance. Identifies coverage gaps and fills them with targeted tests.
Skills are knowledge bases that agents load on demand. They make every agent self-sufficient without requiring access to the Water Framework source code.
| Skill | Purpose |
|---|---|
framework-core-knowledge |
Core philosophy, all base interfaces with full Java signatures, entity hierarchy, Api/SystemApi distinction, QueryBuilder, REST layer, permission system, component lifecycle, anti-patterns |
architecture-knowledge |
DDD patterns, microservice design, module construction, bounded contexts |
authentication-knowledge |
JWT flows, AuthenticationProvider, login/registration, REST security filters |
authorization-knowledge |
@AllowPermissions, @AllowRoles, security interceptors, permission actions, RBAC patterns |
persistence-knowledge |
Repository pattern, JPA implementation, QueryBuilder fluent API, pagination, transactions |
properties-knowledge |
Options pattern, ApplicationProperties, runtime property loading, environment variables |
rest-knowledge |
RestApi hierarchy, JAX-RS / Spring MVC dual support, JWT filters, JSON views, Swagger, exception mapping |
runtime-knowledge |
ComponentRegistry, @FrameworkComponent, @OnActivate/@OnDeactivate, @Inject, OSGi/Spring differences |
karate-testing |
Karate feature file patterns, karate-config.js, test runner configuration, Water test runtime integration |
water-generate |
All yo water:* commands with options, generator-first protocol, .yo-rc.json rules |
test-generation |
Unit test scaffolding, coverage enforcement, SonarQube compliance patterns |
Claude Code coordinates agents automatically. The standard decision tree:
New module / microservice from scratch
└─> water-analyst → water-solution-architect → water-scaffolder → water-backend-architect → water-test-engineer
Add entity to existing project
└─> water-scaffolder (yo water:add-entity) → water-backend-architect → water-test-engineer
Fix bug / implement feature in existing code
└─> water-backend-architect → water-test-engineer
Architecture question / design review
└─> water-solution-architect or water-analyst
Generate/run tests only
└─> water-test-engineer
Each phase validates its output before handing off to the next agent. Sequential phases run one at a time; independent tasks (e.g., scaffolding unrelated entities) run in parallel.
agent-memory/— Per-project agent memory is generated locally at runtime and is not distributed. Each project starts with a clean slate.settings.local.json— Local IDE settings, excluded via.gitignore.- Framework source code — All framework knowledge is embedded in skills; no source access is required.
- Claude Code with
/pluginmarketplace support - Water Framework projects using
yo watergenerator - Java 17+, Gradle
MIT — see LICENSE for details.