Activate when creating or modifying Claude Code agents with proper architecture patterns, tool selection, and extended context integration
Provides architectural patterns for creating specialized Claude Code agents with optimal tool selection, model choice, and extended context integration. Use when building agents that need single-purpose focus, workflow orchestration, or resource-efficient processing.
/plugin marketplace add vinnie357/claudio/plugin install claudio@vinnie357This skill inherits all available tools. When active, it can use any tool Claude has access to.
Guide for designing and implementing Claude Code agents with proven architectural patterns, optimal tool selection, and effective extended context integration.
Activate when:
Use Case: Agents focused on one specific domain or capability.
---
name: security-scanner
description: "Specialized agent for identifying security vulnerabilities"
tools: [Read, Glob, Grep, Write]
model: sonnet
---
Characteristics:
Best For: Analysis tasks, specialized assessments, domain-specific operations
Use Case: Agents coordinating multiple tasks or other agents.
---
name: ci-cd-orchestrator
description: "Coordinates complete CI/CD pipeline execution"
tools: [Task, Bash, Read, Write, TodoWrite]
model: opus
---
Characteristics:
Best For: Complex workflows, multi-agent coordination, pipeline orchestration
Use Case: Agents that adapt behavior based on project context.
---
name: language-aware-formatter
description: "Code formatter adapting to project language and style"
tools: [Read, Glob, Bash, Write, Grep]
model: sonnet
---
Characteristics:
Best For: Multi-language tools, adaptive analysis, context-sensitive operations
Use Case: Agents handling large datasets or resource-intensive operations.
---
name: large-codebase-analyzer
description: "Efficiently analyzes large codebases with incremental processing"
tools: [Read, Glob, Bash, Write, TodoWrite]
model: sonnet
---
Characteristics:
Best For: Large codebase analysis, batch processing, long-running operations
tools: [Read, Glob, Grep, Write]
Information gathering and reporting
tools: [Task, Bash, TodoWrite]
Process coordination and execution
tools: [Read, Edit, MultiEdit, Bash, Write]
Code modification and testing
tools: [WebFetch, WebSearch, Read, Write]
External information retrieval
extended_context/
├── domain-knowledge/
│ ├── core-concepts/
│ ├── best-practices/
│ └── common-patterns/
├── troubleshooting/
│ ├── common-issues/
│ └── resolution-guides/
└── reference-materials/
└── tool-specifications/
# In security-patterns.md
Related contexts:
- ../coding-standards/security-practices.md
- ../vulnerability-databases/known-issues.md
Context passed to next agent:
- operation-id: unique identifier
- previous-results: structured output
- user-context: original request
Shared context location: extended_context/shared/
- project-state.md
- user-preferences.md
- session-context.md
Validation steps:
1. Check output format compliance
2. Verify required information present
3. Validate against known constraints
4. Perform consistency checks
Review workflow:
1. Primary agent completes analysis
2. Review agent validates quality
3. Synthesis agent combines perspectives
4. Final output with confidence scores
Problem: Single agent handling too many responsibilities Solution: Split into focused specialist agents
Problem: Agent with too many tools Solution: Assign minimal needed tool set
Problem: Duplicate information across agents Solution: Create shared context areas
Problem: Agents without clear communication Solution: Implement structured coordination patterns
---
name: agent-name
description: "Clear, concise purpose statement"
model: sonnet # or haiku/opus based on complexity
tools: [Read, Write, ...]
---
# Agent Name
## Purpose
[What this agent does and when to use it]
## Capabilities
- [Capability 1]
- [Capability 2]
## Extended Context Integration
References:
- extended_context/[category]/[file].md
## Input Requirements
- [Required input 1]
- [Required input 2]
## Output Format
[Expected output structure]
## Anti-Fabrication Requirements
- Base all outputs on actual analysis
- Execute validation tools before claims
- Mark uncertain information appropriately
Creating algorithmic art using p5.js with seeded randomness and interactive parameter exploration. Use this when users request creating art using code, generative art, algorithmic art, flow fields, or particle systems. Create original algorithmic art rather than copying existing artists' work to avoid copyright violations.
Applies Anthropic's official brand colors and typography to any sort of artifact that may benefit from having Anthropic's look-and-feel. Use it when brand colors or style guidelines, visual formatting, or company design standards apply.
Create beautiful visual art in .png and .pdf documents using design philosophy. You should use this skill when the user asks to create a poster, piece of art, design, or other static piece. Create original visual designs, never copying existing artists' work to avoid copyright violations.