[feature] Component Builder Skill for AI Coding Agents

## Summary

Create an AI coding agent skill (for Claude Code, Cursor, Codex, etc.) that teaches AI assistants how to build ByteChef components. The skill provides structured, progressive documentation about ByteChef's component architecture, patterns, and conventions so that AI agents can generate production-quality component code that adheres to project standards.

## Problem

Building ByteChef components today requires developers to understand the component handler pattern, action/trigger definitions, connection configurations, the DSL API, testing conventions, and documentation generation. This knowledge is scattered across CLAUDE.md, existing component examples, and tribal knowledge. AI coding agents lack the structured context needed to reliably produce correct, convention-following component code without significant manual guidance.

## Proposed Solution

A skill (hosted in the repository, e.g. `.claude/skills/component-builder/`) that provides AI coding agents with progressive, context-optimized documentation covering:

### Core Knowledge Areas

1. **Component Architecture** — `ComponentHandler` pattern, `@AutoService` annotation, directory structure under `server/libs/modules/components/`
2. **Action Development** — Defining actions with the DSL (`ComponentDsl.action()`), input properties, output schemas, `perform` methods, and connection parameter handling
3. **Trigger Development** — Polling vs. webhook triggers, trigger definitions, listener configurations
4. **Connection Definitions** — OAuth2 flows, API key auth, custom authentication schemes, connection property definitions
5. **Testing Patterns** — Unit test structure, JSON definition auto-generation, `build/resources/test/definition/` cleanup, integration test conventions (`@ComponentIntTest`)
6. **Build & Registration** — Adding to `settings.gradle.kts`, running `./gradlew generateDocumentation`, README.md conventions
7. **SDK & DSL Reference** — Property types (string, integer, bool, object, array, etc.), HTTP client helpers, expression support, dynamic options/properties

### Progressive Disclosure

The skill should load context incrementally rather than dumping everything at once:
- Start with high-level architecture and the scaffolding workflow
- Drill into action/trigger specifics only when the agent is building those
- Surface testing and documentation steps at the appropriate phase

### Agent Workflows

The skill should support end-to-end workflows such as:
- "Create a new component for [Service X] using this OpenAPI spec"
- "Add a new action to the existing [Component Y]"
- "Add webhook trigger support to [Component Z]"
- "Generate tests for [Component W]'s actions"

## Acceptance Criteria

- [ ] Skill is installable and usable by Claude Code (and compatible with Cursor/Codex)
- [ ] Covers component scaffolding, action development, trigger development, connection definitions, testing, and documentation
- [ ] Uses progressive disclosure so agent context is not overwhelmed
- [ ] Generated components follow all ByteChef conventions (Spotless, Checkstyle, naming, DSL patterns)
- [ ] Includes examples derived from existing simple components (e.g., a basic REST API component)
- [ ] Documentation on how to use the skill is included in the repository