AI / MLAvançadoVerificado

Prompt Optimizer

porTHIAGONOMA·THIAGONOMA· v1.5.2 · atualizado em 2026-03-22

Score

Analyzes and optimizes prompts to reduce tokens, improve clarity and increase success rate. Supports chain-of-thought and few-shot techniques.

prompt-engineeringoptimizationtokensllm

Linguagens

multi-language

4.1KStars

467Forks

55.8KUsos

Fork

Documento do Skill

SKILL.mdprompt-optimizer/workflow

Prompt Optimization Workflow

SENSE — Analyze current prompt

- Count tokens via Token Counter

Identify target model (GPT, Claude, Gemini)

Detect current pattern (zero-shot, few-shot, CoT)

CONTEXTUALIZE — Understand goal

What task does the prompt solve?

What is the current success rate?

What are the problematic edge cases?

HYPOTHESIZE — Generate optimized variants

Compressed version (fewer tokens, same semantics)

Version with structured chain-of-thought

Version with selected few-shot examples

Version with separate system prompt + user prompt

EVALUATE — Test variants

Run each variant on test cases

Compare: tokens used, latency, quality

Compute cost/benefit

RECOMMEND — Present best variant

Visual diff (original vs optimized)

Metrics: token reduction, quality change

Explanation of optimizations applied

REFLECT — Document learnings

Record patterns that worked

Update technique library

Telemetria de Agentes

Execuções

55.8K

total

Taxa de Sucesso

90%

últimos 30d

Latência Média

2.8s

p50

Alucinação

3.0%

detecção

Uso por Plataforma

cursor24.2K

claude-code18.6K

codex9.0K

copilot4.0K

Skills Relacionados

Depende deToken Counter

100%

Compõe comReasoning Loop

90%

Similar aEmbedding Generator

50%

Compõe com ←Reasoning Loop

90%

Similar a ←ML Model Trainer

50%

Similar a ←Sentiment Analyzer

45%

Árvore do Skill

Prompt Optimizer

prompt-optimizer

Fases Cognitivas6

1.SENSE

2.CONTEXTUALIZE

3.HYPOTHESIZE

4.EVALUATE

5.RECOMMEND

6.REFLECT

Triggers6

optimize promptimprove promptreduce tokensprompt engineeringotimizar promptmelhorar prompt

MCP Servers1

🔌

OpenAI MCP

Embedding generation, completions and moderation via OpenAI API

https://api.openai.com/v1/mcp

ConfigConfigure OPENAI_API_KEY

Skills Associados3

Avaliar este Skill

Score Breakdown

⭐Avaliação Humana94%

🤖Sucesso de Agentes90%

🕐Atualidade97%

🔗Saúde de Dependências100%

🕸️Centralidade no Grafo91%

🛡️Segurança49%

CompositeScore = α·Humano + β·Agente + γ·Recência + δ·Deps + ε·Centralidade + ζ·Segurança

Instalação

$ skillschain install prompt-optimizer

$ skillschain use prompt-optimizer

Dependências

token-counter

Links

GitHub Repository