senior-prompt-engineer

Production-grade prompt design uses structured architecture patterns like few-shot design and chain-of-thought to reduce hallucinations. This skill provides structured prompts and evaluation workflows to improve accuracy in enterprise chatbots.

You can evaluate RAG pipelines by running the included rag_evaluator.py script with standard Python tooling. It processes input data to measure retrieval and generation performance, outputting structured results for evaluation.

Agent orchestration for LLM systems is best handled through structured system design. The included agent_orchestrator.py script manages task inputs and orchestrates multi-agent workflows, outputting structured results for production deployment.

Yes, the prompt architecture patterns apply to both Claude, GPT-4, and other LLMs. The skill provides generalized prompt engineering structures and evaluation pipelines that integrate with existing workflows across different models.

No external dependencies are required beyond standard Python libraries. The skill relies on standard Python tooling to run the prompt_optimizer, rag_evaluator, and agent_orchestrator scripts within your existing environment.