name: prompt_optimizer
router_kit: AIKit
description: Optimizing token usage and success rate of existing prompts.
metadata:
skillport:
category: ai
tags: [agents, algorithms, artificial intelligence, automation, chatbots, cognitive services, deep learning, embeddings, frameworks, generative ai, inference, large language models, llm, machine learning, model fine-tuning, natural language processing, neural networks, nlp, openai, prompt engineering, prompt optimizer, rag, retrieval augmented generation, tools, vector databases, workflow automation] - prompt-refinement

⚡ Prompt Optimizer

Methodology for increasing prompt efficiency and reducing token cost.

Prompt Optimizer v1.1 - Enhanced

🔄 Workflow

Source: Anthropic - Prompt Engineering Best Practices

Phase 1: Analysis & Benchmarking

• [ ] Analyze: Identify where the current prompt fails or why it consumes too many tokens.
• [ ] Baseline: Record success rate and average token count.

Phase 2: Rewriting & Compression

• [ ] Clarity: Remove unnecessary clauses, use clear imperative moods.
• [ ] Compression: Reduce token count without loss of meaning (Stopwords cleaning etc.).
• [ ] Prompt Chaining: Divide tasks into small, chained prompts instead of a single giant prompt.

Phase 3: Verification (A/B Test)

• [ ] Test: Compare optimized prompt with the old one on same inputs.
• [ ] Score: Score based on response accuracy, speed and cost.

Checkpoints