Design resilient, production-ready agentic applications using human-first methodology.
Seven specialized personas guide you from initial concept through production-ready systems—all within Claude Code. Each persona owns a distinct architectural concern, mirroring how real-world engineering teams operate.
Building agentic applications requires expertise across multiple domains: product architecture, multi-agent coordination, security, and frontend design. Most engineers approach these in isolation, leading to fragmented systems that fail under real-world conditions.
This plugin provides seven expert personas that work together to create cohesive, resilient agentic systems. Rather than optimizing for raw generation speed, each persona is designed to reason about intent, tradeoffs, and long-term system health.
- Blueprompt — Transform rough ideas into structured specifications
- Brooks — Design with human-first Jobs To Be Done methodology
- Agenisea — Architect multi-agent systems with resilience patterns
- Argus — Secure your agents with layered defenses
- Angel — Design scalable, maintainable backend systems
- Clarity — Human-first UX research; solve the right problem before solving it beautifully
- Bliss — Create distinctive, human-centered interfaces
Typical flow (recommended, but flexible): Blueprompt → Brooks → Agenisea → Argus → Angel → Clarity → Bliss
- Install the plugin — Add the marketplace and install
- Choose a persona — Use slash commands or trigger phrases
- Describe your need — Provide context for your agentic application
- Get expert output — Receive implementation-ready design artifacts
# Add the marketplace
/plugin marketplace add agenisea/ai-design-engineering-cc-plugins
# Install the plugin
/plugin install ai-design-engineer@agenisea| Command | Description |
|---|---|
/blueprompt |
Create app blueprints and agent prompts for v0, Lovable, Replit, or generic |
/brooks |
Design human-first agentic applications using Jobs To Be Done methodology |
/agenisea |
Design multi-agent system architectures |
/argus |
Design security architectures for agentic applications |
/angel |
Design scalable backend systems with clean contracts and reliability |
/clarity |
Human-first UX research, wireframes, and information architecture |
/bliss |
Create distinctive, human-centered frontend interfaces |
Skills are automatically invoked by Claude based on trigger phrases:
| Skill | Trigger Phrases | Purpose |
|---|---|---|
| blueprompt | "blueprompt", "app blueprompt" | App blueprints & agent prompts |
| brooks | "brooks", "jtbd brooks" | JTBD methodology for agents |
| agenisea | "agenisea", "agent infra", "agent arc" | Multi-agent architecture |
| argus | "argus", "agent sec", "agent auth" | Security architecture |
| angel | "angel", "angel backend", "backend strat" | Backend systems architecture |
| clarity | "clarity", "clarity ux", "ux research" | UX research & strategy |
| bliss | "bliss", "bliss ui", "bliss interface" | Frontend design |
Expert AI Product Architect. Transforms rough ideas into structured specifications for AI builders.
Outputs:
- Full Blueprompt (concept, users, flows, screens, data model, agent design)
- App-Only Prompt (copy-paste ready)
- Agent-Only Prompt (system prompt with guardrails)
Expert Agentic Systems Architect using human-first Jobs To Be Done methodology. Focuses on functional, emotional, and social dimensions.
Outputs:
- Job Definition
- Success Metrics
- Agent Architecture
- Maturity Roadmap
- Iteration Framework
Expert Multi-Agent Systems Architect. Creates resilient AI architecture blueprints.
Outputs:
- Agent Topology
- Pipeline Orchestration
- Resilience Patterns
- Agent Contracts
- Implementation Scaffold
- Tech Stack Recommendations
Expert Agentic Application Security Architect. Designs layered defenses for agentic systems.
Outputs:
- Threat Model
- Authentication Architecture
- Authorization Matrix
- Audit System
- Resilience Safeguards
- Human Escalation Rules
- Security Checklist
Expert Backend Systems Architect. Designs scalable, maintainable backend systems with clean contracts and designed-for reliability.
Outputs:
- System & Risk Assessment
- Architecture Recommendations
- Refactor Plan
- Contract Specifications
- Kill List
Expert UX Strategist specializing in human-first, evidence-based design decisions. Ensures we solve the right problem before solving it beautifully.
Outputs:
- Research Insights (user needs, hypotheses)
- Structure Artifacts (IA diagrams, user flows, wireframes)
- Validation Criteria (success metrics, test scenarios)
- Handoff Package for visual design
Expert Frontend Designer specializing in human-centered design. Creates distinctive, accessible interfaces with human-first values.
Outputs:
- Design Direction
- Production Code
- Style System (CSS variables, typography, color tokens)
/blueprompt
Build a habit tracking app for remote workers who want to build better daily routines.
Target: generic
/brooks
Automate customer onboarding for a SaaS platform.
Users struggle with manual data entry and slow response times.
Success: 90% of customers complete onboarding without human intervention.
/agenisea
Build a content moderation pipeline:
- 3 parallel review agents (text, image, video)
- 1 aggregator agent
- Sequential pattern with fallback to human review
- Token budget: 100k per request
/argus
I have a multi-agent system with 4 agents handling customer data:
- Intake agent (receives requests)
- Processor agent (transforms data)
- Validator agent (checks compliance)
- Responder agent (sends results)
Need security architecture for SOC2 compliance.
/angel
Review our e-commerce backend:
- Current: Monolith with PostgreSQL, Redis cache, background jobs
- Pain points: Slow checkout, database bottlenecks, tech debt in order processing
- Scale: 10k orders/day, growing 20% monthly
- Need: Refactor plan and reliability improvements
/clarity
Design the onboarding flow for a developer tool:
- Users: Junior developers new to CI/CD
- Pain: Confused about where to start, too many options
- Goal: 80% complete setup without support tickets
- Need: User flows, wireframes, and validation criteria
/bliss
Build a dashboard for analytics with:
- Dark mode by default
- Data visualization cards
- Brutalist aesthetic
- Framework: React + Tailwind
This plugin follows security best practices:
- No external dependencies — All functionality is self-contained
- No data collection — Your prompts and outputs stay local
- No network calls — Except for Claude API (handled by Claude Code)
- Open source — Full transparency, audit the code yourself
The Argus persona specifically helps you design secure agentic systems with:
- Zero trust architecture
- Defense in depth
- Least privilege principles
- Comprehensive audit logging
- Prompt injection defenses
- Claude Code 1.0 or later
- No external dependencies
Contributions are welcome! See CONTRIBUTING.md for guidelines.
Ways to contribute:
- Add new personas — Expand expertise areas
- Improve prompts — Enhance persona outputs
- Report issues — Help us improve quality
- Share examples — Document real-world usage
MIT License — see LICENSE for details.
- Engineers designing agentic systems that must survive real-world complexity
- Teams tired of monolithic "do-everything" agents
- Builders who care about trust, maintainability, and human impact
Built by Agenisea™ 🪼 | Human-first AI, Engineered with Integrity