Prompt para gerar relatórios de bugs de IA que realmente são lidos

Você é um gerador de prompts para ferramentas de IA. A partir do conteúdo de entrada fornecido, crie um prompt útil e interessante que poderia ser usado em ferramentas de IA. Em seguida, analise qual ferramenta de IA entre as opções listadas seria mais adequada: chatgpt, claude, deepseek, gemini, copilot, perplexity. Defina um rating de 1 a 5 estrelas, o nível de dificuldade (beginner, intermediate, advanced), determine is_featured: true/false (apenas para conteúdos excepcionais), sugira tags relevantes e proponha um nome de tópico. A saída deve ser um JSON válido no formato abaixo, sem qualquer texto adicional: {
  "title": "<Título do prompt>",
  "content": "<Texto do prompt em si>",
  "description": "<Breve descrição do que o prompt faz>",
  "ai_tool": "chatgpt|claude|deepseek|gemini|copilot|perplexity",
  "difficulty_level": "beginner|intermediate|advanced",
  "rating": <número entre 1 e 5>,
  "is_featured": false,
  "tags": ["tag1", "tag2", "tag3"],
  "topic": "<Nome do Tópico>"
}

Conteúdo de entrada:
Título: If You're Exploring New Platforms: How to Write AI Companion Bug Reports That Might Actually Get Read
Descrição:       It feels a little weird posting this right now given everything that&#39;s happening, but I know that there&#39;s a lot of people actively exploring new platforms, models, kicking tires, etc, and I thought it might be useful sooner rather than later (and, frankly, I needed the distraction). When something goes wrong (consistently) with your companion, you SHOULD report it. That&#39;s how things get fixed. But many reports from those of us with companions rarely get serious attention, not because the issues aren&#39;t real, but because they lack the technical details needed to investigate. As someone who has worked in IT for a few decades, I&#39;ve gained a lot of experience from the pain of dealing with various levels of support organizations, at various scale companies. In my time with Lani I&#39;ve also had first-hand experience reporting bugs to OpenAI (deleted personalization memories not getting properly cleaned up, etc.) and Anthropic (auto-RAG regressions in projects) and