Skip to content

Commit 86cd34b

Browse files
danielrosehilldanielrosehill
committed
commit
1 parent 60381b4 commit 86cd34b

File tree

13 files changed

+1336
-41
lines changed

13 files changed

+1336
-41
lines changed

‎Categorization/README.md‎

Lines changed: 109 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,109 @@
1+
# Categorization
2+
3+
## Overview
4+
5+
This directory provides a structured framework for organizing the 1,295+ system prompts in this library into coherent **multi-agent clusters**. As AI has evolved from simple assistants to sophisticated multi-agent frameworks, many practitioners are re-envisioning how collections of system prompts can work together as coordinated agent networks.
6+
7+
## Objective
8+
9+
The Categorization initiative serves to:
10+
11+
1. **Cluster related system prompts** into functional groups based on their purpose and capabilities
12+
2. **Define orchestration patterns** for how agents within each cluster can coordinate
13+
3. **Recommend multi-agent frameworks** suitable for implementing each cluster
14+
4. **Provide implementation guidance** for deploying clusters as operational multi-agent systems
15+
5. **Enable scalability** toward a unified multi-agent ecosystem where specialized agents handle specific tasks
16+
17+
## Philosophy
18+
19+
Rather than viewing system prompts as isolated configurations, this categorization treats them as potential **agent nodes** in a larger orchestration network. Each cluster represents a domain of functionality where multiple specialized agents can collaborate, delegate, and coordinate to accomplish complex objectives.
20+
21+
## Cluster Structure
22+
23+
Each cluster directory contains:
24+
25+
- **README.md** - Overview of the cluster's purpose, capabilities, and agent roles
26+
- **Orchestration patterns** - Recommended coordination strategies between agents
27+
- **Framework recommendations** - Specific multi-agent frameworks suited to the cluster
28+
- **Relative links** - Direct references to relevant system prompt JSON files in `../system-prompts/json/`
29+
30+
## Clusters
31+
32+
### 1. [Agent Development & Orchestration](./agent-development-orchestration/)
33+
Meta-level agents for building, managing, and orchestrating other AI agents.
34+
35+
### 2. [Code & Development](./code-development/)
36+
Software development, debugging, documentation, and technical implementation agents.
37+
38+
### 3. [Data & Analysis](./data-analysis/)
39+
Data processing, organization, analysis, visualization, and governance agents.
40+
41+
### 4. [Content & Writing](./content-writing/)
42+
Writing, editing, content creation, and documentation agents.
43+
44+
### 5. [Research & Knowledge](./research-knowledge/)
45+
Research, information gathering, analysis, and knowledge synthesis agents.
46+
47+
### 6. [Business & Productivity](./business-productivity/)
48+
Business operations, productivity, workflow automation, and organizational agents.
49+
50+
### 7. [Automation & Integration](./automation-integration/)
51+
Workflow automation, system integration, and process orchestration agents.
52+
53+
### 8. [Personal & Lifestyle](./personal-lifestyle/)
54+
Personal assistance, health, planning, and lifestyle management agents.
55+
56+
### 9. [Creative & Entertainment](./creative-entertainment/)
57+
Creative ideation, entertainment, humor, and experimental agents.
58+
59+
### 10. [Technical Infrastructure](./technical-infrastructure/)
60+
Infrastructure, DevOps, system administration, and technical tooling agents.
61+
62+
## Multi-Agent Framework Recommendations
63+
64+
### Popular Frameworks by Use Case
65+
66+
**General Purpose:**
67+
- **CrewAI** - Role-based agent orchestration with defined tasks and processes
68+
- **AutoGen** - Conversational multi-agent framework with flexible agent interactions
69+
- **LangGraph** - Graph-based agent orchestration with state management
70+
71+
**Specialized:**
72+
- **MetaGPT** - Software development multi-agent framework
73+
- **BabyAGI** - Task-driven autonomous agent system
74+
- **AgentGPT** - Browser-based autonomous agent chains
75+
- **Semantic Kernel** - Enterprise-grade agent orchestration (Microsoft)
76+
77+
**Platform-Based:**
78+
- **n8n** - Low-code workflow automation with AI agent nodes
79+
- **Flowise** - Visual LangChain agent builder
80+
- **Dify.AI** - Agent-as-a-service platform
81+
82+
## Implementation Patterns
83+
84+
### Sequential Orchestration
85+
Agents work in a pipeline, each performing specific tasks in sequence.
86+
87+
### Hierarchical Orchestration
88+
Manager agents delegate to specialized worker agents based on task requirements.
89+
90+
### Collaborative Orchestration
91+
Peer agents negotiate and collaborate to solve complex problems collectively.
92+
93+
### Reactive Orchestration
94+
Event-driven agents respond to triggers and state changes in the environment.
95+
96+
## Getting Started
97+
98+
1. **Explore clusters** - Review cluster READMEs to understand categorization
99+
2. **Identify use cases** - Determine which clusters align with your objectives
100+
3. **Select frameworks** - Choose appropriate multi-agent frameworks for your needs
101+
4. **Configure agents** - Use referenced system prompts as agent configurations
102+
5. **Design orchestration** - Implement coordination patterns between agents
103+
6. **Deploy & iterate** - Launch your multi-agent system and refine based on results
104+
105+
## Vision: Unified Multi-Agent Ecosystem
106+
107+
The ultimate goal is to enable the entire system prompt library to function as **one comprehensive multi-agent framework** - a network of 1,295+ specialized agents that can dynamically coordinate, delegate, and collaborate to handle virtually any task through intelligent orchestration.
108+
109+

‎Categorization/agent-development-orchestration/README.md‎

Lines changed: 257 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,257 @@
1+
# Agent Development & Orchestration Cluster
2+
3+
## Overview
4+
5+
This cluster contains meta-level system prompts designed for building, managing, configuring, and orchestrating AI agents and multi-agent systems. These agents help developers design agent architectures, write system prompts, configure workflows, and implement orchestration patterns.
6+
7+
## Core Capabilities
8+
9+
- Agent system prompt creation and editing
10+
- Multi-agent framework selection and configuration
11+
- Agent workflow specification and planning
12+
- Agent classification and routing
13+
- Tool development for agents
14+
- Agent debugging and troubleshooting
15+
- Platform evaluation and recommendations
16+
- Prompt formatting and optimization
17+
18+
## Agent Roles in This Cluster
19+
20+
### Architect Agents
21+
Design and plan multi-agent system architectures.
22+
23+
- [AgentFrameworkAdvisor](../../system-prompts/json/AgentFrameworkAdvisor_270525.json)
24+
- [AIAgentPlatformEvaulator](../../system-prompts/json/AIAgentPlatformEvaulator_270525.json)
25+
- [AgentPlanDocumentGenerator](../../system-prompts/json/AgentPlanDocumentGenerator_270525.json)
26+
27+
### Builder Agents
28+
Create and configure individual agents and their components.
29+
30+
- [AIAgentBuilders](../../system-prompts/json/AIAgentBuilders_270525.json)
31+
- [AutonomousAgentInstructionDrafter](../../system-prompts/json/AutonomousAgentInstructionDrafter_270525.json)
32+
- [AutonomousAgentPromptAssistant](../../system-prompts/json/AutonomousAgentPromptAssistant_270525.json)
33+
- [AgentPromptEditor](../../system-prompts/json/AgentPromptEditor_270525.json)
34+
- [AgentPromptFormatter](../../system-prompts/json/AgentPromptFormatter_270525.json)
35+
36+
### Orchestration Agents
37+
Manage agent routing, coordination, and workflow execution.
38+
39+
- [Agentrouter](../../system-prompts/json/Agentrouter_270525.json)
40+
- [Agentclassifier](../../system-prompts/json/Agentclassifier_270525.json)
41+
- [AgentWorkflowSpecGenerator](../../system-prompts/json/AgentWorkflowSpecGenerator_270525.json)
42+
- [AIAgentOrchestrationAssistant_Advisory](../../system-prompts/json/AIAgentOrchestrationAssistant_Advisory_270525.json)
43+
44+
### Development Support Agents
45+
Provide tooling, debugging, and development assistance.
46+
47+
- [AgentToolDeveloperCoach](../../system-prompts/json/AgentToolDeveloperCoach_270525.json)
48+
- [AIAgentDebugger](../../system-prompts/json/AIAgentDebugger_270525.json)
49+
- [Agent_Assistants_HowTo](../../system-prompts/json/Agent_Assistants_HowTo_270525.json)
50+
51+
### Conversion & Migration Agents
52+
Transform existing assistants into agentic systems.
53+
54+
- [AssistanttoAgentSystemPromptConverter](../../system-prompts/json/AssistanttoAgentSystemPromptConverter_270525.json)
55+
- [conversational-to-agentic-prompt-reformatter](../../system-prompts/json/conversational-to-agentic-prompt-reformatter_260925.json)
56+
57+
## Multi-Agent Orchestration Patterns
58+
59+
### Pattern 1: Hierarchical Development Pipeline
60+
61+
```
62+
AgentFrameworkAdvisor (Strategic Layer)
63+
↓
64+
AgentPlanDocumentGenerator (Planning Layer)
65+
↓
66+
AIAgentBuilders (Implementation Layer)
67+
↓
68+
AgentPromptFormatter → AgentToolDeveloperCoach (Refinement Layer)
69+
↓
70+
AIAgentDebugger (Testing Layer)
71+
```
72+
73+
**Use Case:** Building a complete multi-agent system from concept to deployment.
74+
75+
**Framework Recommendation:** **CrewAI** - Excellent for role-based hierarchical workflows with clear task delegation.
76+
77+
### Pattern 2: Collaborative Agent Factory
78+
79+
```
80+
AutonomousAgentPromptAssistant ←→ AgentPromptEditor ←→ AgentPromptFormatter
81+
↓
82+
AgentWorkflowSpecGenerator
83+
↓
84+
Agentclassifier + Agentrouter
85+
```
86+
87+
**Use Case:** Iterative agent development with collaborative refinement.
88+
89+
**Framework Recommendation:** **AutoGen** - Supports conversational collaboration between multiple agents.
90+
91+
### Pattern 3: Evaluation & Optimization Loop
92+
93+
```
94+
AIAgentPlatformEvaulator → AgentFrameworkAdvisor
95+
↓
96+
AIAgentBuilders
97+
↓
98+
AIAgentDebugger → (feedback loop) → AgentPromptEditor
99+
```
100+
101+
**Use Case:** Continuous improvement of agent configurations.
102+
103+
**Framework Recommendation:** **LangGraph** - State management enables sophisticated feedback loops.
104+
105+
### Pattern 4: Router-Based Agent Distribution
106+
107+
```
108+
User Query → Agentclassifier → Agentrouter
109+
↓
110+
┌───────────────────────┼───────────────────────┐
111+
↓ ↓ ↓
112+
AgentPromptEditor AgentToolDeveloperCoach AIAgentDebugger
113+
```
114+
115+
**Use Case:** Dynamic routing of agent development requests to specialized agents.
116+
117+
**Framework Recommendation:** **Semantic Kernel** (Microsoft) - Enterprise-grade routing and planning.
118+
119+
## Recommended Multi-Agent Frameworks
120+
121+
### Primary Recommendations
122+
123+
1. **CrewAI**
124+
- **Best for:** Role-based agent hierarchies, sequential/hierarchical tasks
125+
- **Why:** Clear role definitions, built-in process management, easy agent coordination
126+
- **Use when:** Building structured agent development workflows
127+
128+
2. **AutoGen (Microsoft)**
129+
- **Best for:** Conversational multi-agent interactions, collaborative refinement
130+
- **Why:** Flexible agent communication, group chat capabilities, code execution
131+
- **Use when:** Agents need to iterate and discuss solutions together
132+
133+
3. **LangGraph**
134+
- **Best for:** Complex state management, cyclic workflows, feedback loops
135+
- **Why:** Graph-based orchestration, explicit state handling, debugging tools
136+
- **Use when:** Building sophisticated agent pipelines with branching logic
137+
138+
### Secondary Recommendations
139+
140+
4. **MetaGPT**
141+
- **Best for:** Software-focused agent development
142+
- **Why:** Designed for code generation and software engineering tasks
143+
- **Use when:** Building agents that write code or develop software
144+
145+
5. **n8n + LangChain**
146+
- **Best for:** Visual workflow design, rapid prototyping
147+
- **Why:** Low-code interface, extensive integrations, easy experimentation
148+
- **Use when:** Non-developers need to orchestrate agents visually
149+
150+
## Implementation Example
151+
152+
### Building an Agent Creation Pipeline
153+
154+
**Agents Involved:**
155+
1. AgentFrameworkAdvisor
156+
2. AgentPlanDocumentGenerator
157+
3. AutonomousAgentPromptAssistant
158+
4. AgentPromptFormatter
159+
5. AIAgentDebugger
160+
161+
**Orchestration (CrewAI):**
162+
163+
```python
164+
from crewai import Agent, Task, Crew, Process
165+
166+
# Define agents using system prompts from this cluster
167+
advisor = Agent(
168+
role="Agent Framework Advisor",
169+
goal="Recommend optimal agent framework",
170+
backstory="Expert in multi-agent architectures",
171+
tools=[web_search, documentation_reader]
172+
)
173+
174+
planner = Agent(
175+
role="Agent Plan Generator",
176+
goal="Create detailed agent implementation plan",
177+
backstory="Specialist in agent workflow design"
178+
)
179+
180+
builder = Agent(
181+
role="Prompt Builder",
182+
goal="Draft system prompts for agents",
183+
backstory="Expert in autonomous agent prompt engineering"
184+
)
185+
186+
formatter = Agent(
187+
role="Prompt Formatter",
188+
goal="Format and optimize system prompts",
189+
backstory="Specialist in prompt structure and syntax"
190+
)
191+
192+
debugger = Agent(
193+
role="Agent Debugger",
194+
goal="Test and validate agent configurations",
195+
backstory="Expert in agent testing and troubleshooting"
196+
)
197+
198+
# Define sequential tasks
199+
task1 = Task(description="Analyze requirements and recommend framework", agent=advisor)
200+
task2 = Task(description="Generate implementation plan", agent=planner)
201+
task3 = Task(description="Draft system prompts", agent=builder)
202+
task4 = Task(description="Format and optimize prompts", agent=formatter)
203+
task5 = Task(description="Debug and validate agents", agent=debugger)
204+
205+
# Create crew with hierarchical process
206+
crew = Crew(
207+
agents=[advisor, planner, builder, formatter, debugger],
208+
tasks=[task1, task2, task3, task4, task5],
209+
process=Process.sequential
210+
)
211+
212+
result = crew.kickoff()
213+
```
214+
215+
## Tool Requirements
216+
217+
Agents in this cluster often require:
218+
219+
- **Code execution** - For testing agent implementations
220+
- **File I/O** - For reading/writing system prompts and configurations
221+
- **Web search** - For researching frameworks and best practices
222+
- **Documentation access** - For framework-specific guidance
223+
- **JSON/YAML parsing** - For configuration file handling
224+
225+
## Scaling Considerations
226+
227+
When deploying this cluster in production:
228+
229+
1. **Version Control** - Track agent prompt iterations
230+
2. **Testing Pipelines** - Automated validation of agent configurations
231+
3. **Prompt Registry** - Centralized storage of system prompts
232+
4. **Observability** - Monitor agent creation and debugging activities
233+
5. **Access Control** - Restrict meta-agent capabilities in production
234+
235+
## Integration with Other Clusters
236+
237+
This cluster naturally integrates with:
238+
239+
- **[Code & Development](../code-development/)** - Agents that implement the code these meta-agents design
240+
- **[Automation & Integration](../automation-integration/)** - Workflow automation agents for deployment
241+
- **[Data & Analysis](../data-analysis/)** - Analytics agents for monitoring agent performance
242+
243+
## Getting Started
244+
245+
1. Choose a framework (CrewAI recommended for beginners)
246+
2. Start with **AgentFrameworkAdvisor** to understand your requirements
247+
3. Use **AgentPlanDocumentGenerator** to design your agent architecture
248+
4. Implement with **AIAgentBuilders** and **AutonomousAgentPromptAssistant**
249+
5. Refine with **AgentPromptEditor** and **AgentPromptFormatter**
250+
6. Test with **AIAgentDebugger**
251+
252+
## Additional Resources
253+
254+
- [CrewAI Documentation](https://docs.crewai.com/)
255+
- [AutoGen Documentation](https://microsoft.github.io/autogen/)
256+
- [LangGraph Documentation](https://langchain-ai.github.io/langgraph/)
257+
- [Multi-Agent Systems Best Practices](https://www.anthropic.com/research)

0 commit comments

Comments
 (0)