Pre-Made Prompts for AI Coding Assistants

Copy these prompts into Claude Code, ChatGPT Codex, Cursor, or any AI coding assistant to integrate PyRapide into your project. Each prompt references the actual INTEGRATION.md file in the PyRapide repository so your AI assistant has full context.

These prompts are designed to give your AI assistant everything it needs to perform the integration correctly on the first try. They reference the specific adapter classes, integration patterns, and constraint patterns documented in the PyRapide codebase.

Universal Integration Prompt

I want to add PyRapide causal event tracing to my existing agentic AI project. 
PyRapide is installed via `pip install pyrapide`. 

First, identify which agentic framework my project uses (AutoGen, LangGraph, CrewAI, LlamaIndex, Agno, OpenAI Swarm, MetaGPT, or FlowiseAI).

Then read the integration guide at `pyrapide/agent_templates/{framework}/INTEGRATION.md` in the PyRapide package.

Add the appropriate adapter with these steps:
1. Import the adapter class
2. Create the adapter instance
3. Hook it into my framework (using the pattern from the INTEGRATION.md)
4. Add a StreamProcessor with constraint enforcement
5. Add post-run analysis using pyrapide.queries

Use the AgentPatterns library for pre-built patterns:
- AgentPatterns.tool_roundtrip() for tool call completion tracking
- AgentPatterns.any_error() for error detection
- AgentPatterns.full_tool_lifecycle() for end-to-end tracing

Do not modify my existing framework code beyond the 1-3 lines needed for integration.

Per-Framework Prompts

Use the prompt specific to your framework for the most targeted integration:

AutoGen Prompt

I'm using Microsoft AutoGen for my multi-agent project. Add PyRapide causal 
tracing by reading `pyrapide/agent_templates/autogen/INTEGRATION.md` and adding 
the AutoGenAdapter to my project. Hook into AutoGen's runtime event system using 
`adapter = AutoGenAdapter("my_app", runtime=runtime)`. Set up constraints to 
ensure every tool call completes with a result event using 
AgentPatterns.tool_roundtrip(). Add backward_slice analysis after each 
conversation to trace agent decision chains.

Framework Detection Prompt

Analyze my project and identify which agentic AI framework(s) I'm using. Check 
for imports from: autogen, langchain/langgraph, crewai, llama_index, agno/phidata, 
swarm, metagpt, or flowise API calls. Then recommend which PyRapide agent template 
adapter to use and show me the integration code.