Anthropic Official Skills and Plugins¶
STATUS: active OWNER: infrastructure CATEGORY: tooling COST: free (all skills and plugins are local) ADDED: 2026-04-01 UPDATED: 2026-04-01
OVERVIEW¶
GE uses 6 Anthropic-compatible skills and 8 official plugins installed at the project level. Skills provide domain-specific knowledge injection. Plugins provide structured workflows and language tooling. Both auto-activate based on context — no manual triggering needed for most.
SKILLS (in .claude/skills/)¶
| Skill | Purpose | Used By | Source |
|---|---|---|---|
| ui-ux-pro-max | Design intelligence — 67 styles, 161 palettes, typography, UX guidelines | Floris, Floor, Martijn, Alexander | nextlevelbuilder/ui-ux-pro-max-skill |
| nano-banana | Gemini image generation — mockups, hero images, brand exploration | Felice, Christel, Ilian, Alexander, Valentijn | kingbootoshi/nano-banana-2-skill |
| webapp-testing | Playwright-based web app testing patterns | Marije, Judith | anthropics/skills |
| brand-guidelines | Structured brand voice enforcement | Rick, Christel, Valentijn, Margot | anthropics/skills |
| mcp-builder | Accelerates MCP server creation | Infrastructure/tooling | anthropics/skills |
| skill-creator | Accelerates skill authoring | Infrastructure/tooling | anthropics/skills |
SKILL DETAILS¶
ui-ux-pro-max — The primary design intelligence layer. Provides 67 named design styles, 161 curated color palettes, typography pairing rules, and UX interaction guidelines. Auto-activates on any UI-related task. See also: ui-ux-pro-max.md.
nano-banana — Gemini-powered image generation for mockups, hero images, and brand visual exploration. Uses Flash model for cheap iteration, Pro model for final deliverables. See also: nano-banana-2.md.
webapp-testing — Provides Playwright-based testing patterns for web application E2E and integration tests. Used by Marije and Judith during Phase 8 (Integration) testing.
brand-guidelines — Enforces structured brand voice across all content-producing agents. Ensures tone, terminology, and messaging consistency with the client's approved brand identity.
mcp-builder — Accelerates creation of new MCP (Model Context Protocol) servers. Used by infrastructure and tooling agents when extending GE's tool ecosystem.
skill-creator — Accelerates authoring of new Claude skills. Used when GE needs to package domain knowledge as a reusable skill.
PLUGINS (in .claude/plugins/)¶
| Plugin | Purpose | Used By | Source |
|---|---|---|---|
| feature-dev | Structured feature development workflow (most popular, 89k installs) | All dev agents (Urszula, Floris, Floor, Maxim, Martijn) | anthropics/claude-plugins-official |
| code-review | Automated code review with specialized agents | Koen | anthropics/claude-plugins-official |
| security-guidance | Security warnings when editing sensitive files | Victoria, all dev agents | anthropics/claude-plugins-official |
| code-simplifier | Code simplification and refactoring | All dev agents, Koen | anthropics/claude-plugins-official |
| typescript-lsp | TypeScript language server (go-to-def, find refs, errors) | Floris, Floor, Urszula, Maxim | anthropics/claude-plugins-official |
| pyright-lsp | Python type checking | Koen, orchestrator work | anthropics/claude-plugins-official |
| swift-lsp | Swift language server for iOS | Martijn | anthropics/claude-plugins-official |
| pr-review-toolkit | PR review agents for comments, tests, error handling | Marta, Iwona | anthropics/claude-plugins-official |
PLUGIN DETAILS¶
feature-dev — Enforces a structured feature development workflow: requirements analysis, implementation plan, coding, testing, documentation. All dev agents use this as their primary development loop.
code-review — Provides automated code review with specialized review agents for different concerns (logic, style, security, performance). Primary tool for Koen's deterministic quality gate in Phase 8.
security-guidance — Surfaces security warnings when agents edit sensitive files (auth, crypto, secrets, config). Active for Victoria's threat modeling and as a passive guard for all dev agents.
code-simplifier — Identifies opportunities for code simplification and refactoring. Helps enforce the 300-line file limit from CODEBASE-STANDARDS.md. Used by all dev agents and Koen during review.
typescript-lsp — Provides TypeScript language server capabilities: go-to-definition, find references, real-time type errors. Essential for Floris, Floor, Urszula, and Maxim during web frontend development.
pyright-lsp — Python type checking via Pyright. Used by Koen for Python code review and during orchestrator/executor development.
swift-lsp — Swift language server for iOS development. Used exclusively by Martijn for native iOS (Swift/SwiftUI) work.
pr-review-toolkit — PR review agents that analyze comments, test coverage, and error handling. Used by Marta and Iwona at the merge gate (Phase 8, G8).
INSTALLATION¶
Skills and plugins are installed at the project level in .claude/skills/ and .claude/plugins/. They auto-activate based on context — no manual triggering needed for most.
To add a new skill:
To add a new plugin:
LSP BINARIES¶
The LSP plugins require language server binaries:
| Binary | Install Method | Notes |
|---|---|---|
| typescript-language-server | npm install -g typescript-language-server typescript |
Installed globally on all dev environments |
| pyright | pip install pyright |
Installed in executor Python environment |
| swift-lsp | Requires Xcode | macOS only — Martijn's local environment |
CROSS_REFERENCES¶
- UI UX Pro Max integration: ui-ux-pro-max.md
- Nano Banana 2 integration: nano-banana-2.md
- Playwright integration: playwright.md
- Design tool workflow: design-tool-workflow.md
- Testing standards: testing.md
- Delivery swimlanes: delivery-swimlanes.md