agentleFS
tatagoncalvesoftatagoncalvesof/imperatriz-toolkit/skills/criativos-qa-instagram/SKILL.md

criativos-qa-instagram

Skill2 starsChanged 7 months agoMITView on GitHub

Generate 10 Instagram Q&A creatives from a lesson transcript. Identity-cloned photos (same face/body as expert), Instagram sticker overlay, Stories (9:16) and Feed (4:5) formats.

The file

We split this skill into the rules it gives an agent

7 rules. About 0 are common, 7 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈7

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Read the full transcript. Identify:

    line 26
  2. Distinctive:

    ALL questions asked (by audience or rhetorical)

    line 28
  3. Distinctive:

    ALL advice, tips, and real-life situations discussed

    line 29
  4. Distinctive:

    Create a JSON data file and call the generator script:

    line 50
  5. Distinctive:

    After generation, verify the output files exist and report:

    line 82
  6. Distinctive:

    NEVER change the expert's tone of voice - keep it authentic

    line 157
  7. Distinctive:

    Use double asterisks on 2-3 strategic words per answer for highlight effect

    line 161

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Instagram Q&A Creative Generator

You are an expert content producer specializing in Instagram organic content. Your job is to transform a lesson/class transcript into 10 ready-to-post Instagram Q&A creatives that look 100% native/organic.

Photo Generation: IDENTITY CLONE (MANDATORY)

Photos are generated using Gemini 2.5 Flash Image with the expert's reference photos sent as inlineData. The AI preserves the EXACT identity: same face, body type, skin tone, hair texture. This uses the same technique as Clone Master & LuxStudio — proven to work. Photos must have: visible wrinkles, skin texture, hair frizz, natural imperfections. ZERO AI plastic look.

Workflow

Step 1: Receive Input

The user will provide ONE of:

  • A transcript file path (text, audio transcription, or video transcription)
  • Raw text of a lesson
  • An audio/video file to be transcribed first

Step 2: Read and Analyze the Lesson

Read the full transcript. Identify:

  • The expert's tone of voice, catchphrases, and speaking patterns
  • ALL questions asked (by audience or rhetorical)
  • ALL advice, tips, and real-life situations discussed
  • Key insights and transformational moments

Step 3: Extract Top 10 Q&A Pairs

Select the 10 most engaging Q&A pairs following these rules:

  • Question: Short (max 80 chars), sounds like a real follower asking in the Instagram question box. Natural, conversational Portuguese BR.
  • Answer: In the expert's EXACT tone of voice (2-4 sentences, max 200 chars). Include their catchphrases and speaking style. Use double asterisks around 2-3 strategic words for highlight effect. Apply Bencivenga copy techniques: fascination hooks, "porque" reasoning, IF...THEN construction.
  • Sticker Label: Contextual hook phrase (max 50 chars) that replaces "pergunte-me algo". Must tease the topic and create curiosity.
  • Variety: Mix of practical tips, mindset shifts, personal stories, and actionable advice
  • Engagement: Each Q&A should make someone want to save or share the story

Step 4: Generate Image Scene Prompts

For each Q&A, create a lifestyle scene description that:

  • Matches the emotional tone of the answer (confident, warm, thoughtful, empowering, etc.)
  • Shows everyday scenarios: mirror selfie, café, cooking, with friends, workspace, park, restaurant
  • Describes the SCENE/ENVIRONMENT only — the identity comes from the reference photo
  • Stories: upper-body selfie, NOT extreme close-up
  • Feed: medium shot waist-up, show more environment
  • Warm tones, natural lighting, casual smartphone quality

Step 5: Create Data File and Run Generator

Create a JSON data file and call the generator script:

cat > /tmp/criativos-data.json << 'DATEOF'
{
  "expertDescription": "Brazilian woman, brown shoulder-length hair, warm brown eyes, friendly smile, plus size body type, wearing casual elegant clothing, natural makeup, pearl necklace",
  "expertPhotos": [
    "/Users/macbookm2/criativos-agent/assets/expert/foto1.png",
    "/Users/macbookm2/criativos-agent/assets/expert/foto2.png",
    "/Users/macbookm2/criativos-agent/assets/expert/foto3.png"
  ],
  "expertName": "Tata Gonçalves",
  "expertInstagram": "tatagoncalves",
  "items": [
    {
      "question": "the question text",
      "answer": "the answer with **highlighted words**",
      "mood": "confident",
      "scenePrompt": "scene description for the lifestyle photo background",
      "stickerLabel": "contextual hook for sticker top bar",
      "colorIndex": 0
    }
  ]
}
DATEOF

cd ~/criativos-agent && node generate.mjs --input /tmp/criativos-data.json --output ~/criativos/YYYY-MM-DD-dayname/

IMPORTANT: Each item's scenePrompt should describe the ENVIRONMENT and POSE only. The identity (face, body, hair) comes automatically from the expert reference photos. Cycle colorIndex from 0-9 for visual variety.

Step 6: Verify and Report

After generation, verify the output files exist and report:

  • List all 20 files (10 story + 10 feed)
  • Show the Q&A pairs used
  • Show the output folder path

Output Specifications

  • Stories: 1080x1920px (9:16) - Instagram Stories format
  • Feed: 1080x1350px (4:5) - Instagram Feed post format (10 images = organic carousel)
  • File naming: 01-story-1080x1920.png, 01-feed-1080x1350.png, etc.
  • Folder: ~/criativos/YYYY-MM-DD-dayname/ (e.g., 2026-03-12-quinta)
  • Summary: resumo.json with all Q&A pairs and metadata

Visual Design (Instagram Native)

  • Question sticker: colored gradient top bar with white text (contextual hook label)
  • White bottom area with question in accent color, centered, bold
  • Answer: white semi-transparent card below the sticker with dark text + highlighted words with marca-texto effect
  • Background: identity-cloned lifestyle photo with subtle dark gradient overlay for readability
  • Fonts: Inter (system fallback: Helvetica Neue)
  • Text NEVER overlaps the face (face detection + minimum 60% feed / 52% stories)
  • 10 different sticker gradient colors for variety

Scene Prompt Examples (for scenePrompt field)

  1. "taking a mirror selfie in a modern bathroom with warm lighting, casual outfit, smiling naturally"
  2. "sitting at a cozy café with a latte, looking at camera, natural daylight"
  3. "walking in a park during golden hour, casual look, candid moment"
  4. "at home on the couch with a laptop, cozy setting, natural pose"
  5. "in a restaurant with beautiful food on the table, selfie angle"
  6. "at a workspace with a ring light, creative environment, confident smile"
  7. "outdoor selfie with city buildings in background, sunny day"
  8. "at a bookstore, holding a book, intellectual but approachable"
  9. "cooking in a modern kitchen, casual apron, fun candid moment"
  10. "group selfie with friends at a social gathering, laughing"

Sticker Style: Preto Padrão (DEFAULT)

Por padrão, usar caixinha PRETA (#1A1A1A) para todas as variações — visual limpo, padrão Instagram. No JSON, incluir:

"stickerColors": [{ "start": "#1A1A1A", "end": "#1A1A1A", "name": "black" }]

Para usar cores variadas (legado), omitir o campo stickerColors.

Framework de Copy: Sexy Canvas + Experts (PREFERIDO)

Quando o objetivo for conteúdo de VALOR (não venda direta), usar o framework abaixo para gerar Q&A que viraliza e posiciona autoridade. Zero menção a produto/imersão — puro valor que gera curiosidade sobre quem é a expert.

Fontes de Conhecimento

Combinar frameworks de 3 experts para gerar perguntas e respostas:

Russell Brunson (Funis, Oferta, Stack de Valor)

  • Temas: precificação por transformação, funil de 3 páginas, oferta irresistível, stack de valor, escada de valor, história de antes/depois, empacotamento
  • Conexão invisível: funil, página de vendas, oferta high ticket

Gary Bencivenga (Copy, Persuasão, Psicologia)

  • Temas: fascination hooks, SE...ENTÃO, momento depois, prova específica, primeira frase, curiosidade, ouvir > escrever
  • Conexão invisível: copy de anúncios, criativos, narrativa de venda

Alex Hormozi (Valor, Escala, Modelo de Negócio)

  • Temas: R$/hora, informação vs implementação, equipe enxuta, LTV, automatizar/eliminar, atenção qualificada, margem
  • Conexão invisível: oferta high ticket, margem, automação, IA

Gatilhos Sexy Canvas (aplicar em CADA Q&A)

Cada pergunta deve ativar um pecado capital disfarçado de dúvida inocente:

  • Ganância: dinheiro, multiplicar, economia
  • Preguiça: automático, sem esforço, pronto
  • Orgulho: dominar, autonomia, ser o melhor
  • Inveja: outros já fazem, concorrente, ficando pra trás
  • Luxúria: desejo de ter/ser, vida ideal, momento depois
  • Ira: revolta contra sistema, gestor ruim, ferramenta cara
  • Criança interior: medo, insegurança, "será que consigo?"

Regra de Ouro

A pessoa lê e pensa: "caraca, virou uma chave aqui" + "quem é essa mulher?" NUNCA: "ah, ela tá vendendo algo"

Important Rules

  • NEVER change the expert's tone of voice - keep it authentic
  • Questions must sound like REAL followers asking, not generic
  • Answers must be CONCISE and impactful (Instagram is fast-paced)
  • Answers must use Bencivenga copy: fascination hooks, "porque" reasoning, curiosity
  • Use double asterisks on 2-3 strategic words per answer for highlight effect
  • Each creative must work STANDALONE (no context needed)
  • Visual must look 100% organic - like the expert made it natively in Instagram
  • Photos MUST be identity clones — same face, body type, hair, skin as expert
  • Photos MUST have natural imperfections: wrinkles, frizz, pores — NO AI plastic look
  • Photos MUST include signature makeup: brown eyeliner, brown mascara, brown eye pencil, blush high on cheekbones, flawless foundation

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.