Communitygithub.com

builtnotboosted/configuración-de-marca

Incorporación basada en entrevistas que construye el espacio de trabajo de marketing que leen todas las demás habilidades de Built Not Boosted. Hace preguntas en lotes pequeños (con opciones múltiples cuando es posible), redacta respuestas a partir del material que el usuario ya tiene (sitio web, biografías, publicaciones pasadas) y luego escribe brand/brand-profile.md junto con archivos iniciales de aprendizaje de voz, aprendizaje de rendimiento, fuentes y una carpeta de imágenes de referencia. Vuelve a ejecutarlo para actualizar una sola sección. Úsalo cuando alguien diga "configura mi marca", "incorpórame", "crea mi perfil de marca", "configura el espacio de trabajo de marketing", "inicia el plugin", "actualiza mi perfil de marca", "cambia mi sección de voz/plataformas/cumplimiento normativo", o cuando otra habilidad no encuentre el espacio de trabajo de marketing.

¿Qué es configuración-de-marca?

Esta habilidad establece el espacio de trabajo de marketing y su fuente de verdad central, `brand/brand-profile.md`, que es leído por todas las demás habilidades del paquete (como escritura, investigación, programación, datos de plataforma, guion de video, generación de imágenes, edición de video, bucle de revisión y bucle de voz) para conocer las constantes de la marca: quién es el usuario, a quién sirve, cómo suena, dónde publica, qué nunca dice y qué líneas de cumplimiento normativo aplican. El objetivo es hacerlo bien una vez para que cada pieza de contenido comience más cerca de estar terminada. La habilidad realiza el proceso de incorporación a través de una entrevista en lotes pequeños de preguntas, ofreciendo opciones de selección múltiple cuando es posible, y redacta respuestas a partir del material existente del usuario como su sitio web, biografías o publicaciones pasadas. Soporta una configuración completa (de 10 a 15 minutos) o una configuración rápida (de 5 minutos) que cubre los elementos esenciales. La habilidad se ejecuta en plataformas como Cursor, Windsurf, Lovable, Cline y cualquier entorno compatible con el protocolo Model Context Protocol (MCP), y es particularmente útil para actualizar secciones específicas del perfil de marca sin sobrescribir todo el documento, asegurando que los encabezados exactos del esquema se mantengan para que otras habilidades puedan leerlos correctamente.

Compatible con~Claude Code~Codex CLI~Cursor
npx skills add https://github.com/builtnotboosted/built-not-boosted-skills/tree/HEAD/plugins/built-not-boosted/skills/brand-setup

Preguntar en tu IA favorita

Abre un nuevo chat con esta habilidad de agente ya precargada.

Documentación

Brand setup

This skill creates the marketing workspace and its source of truth, brand/brand-profile.md. Every other skill in the pack (writing, research, scheduling, platform-data, video-scripting, image-generation, video-editing, review-loop, voice-loop) reads that file for the brand constants: who you are, who you serve, how you sound, where you post, what you never say and which compliance lines apply. Get it right once and every piece of content starts closer to finished.

It owns the interview, the workspace layout and the profile schema. It does not write content, choose formats or decide strategy beyond what the user tells it.

Before you start

  1. Look for an existing workspace. Default is marketing/ in the working directory; the user may name another folder.
  2. If marketing/brand/brand-profile.md already exists, switch to re-run mode (below). Never overwrite a filled profile wholesale.
  3. Open references/brand-profile-template.md. Its headings are the exact schema the other skills read, so the output must keep them word for word.
  4. Glance at references/worked-example.md if you want to see what a finished profile looks like at the right level of detail.

The standard

  • Exact headings. The twelve ## headings in the template, in that order, spelled exactly. Other skills look them up by name, so a renamed heading is a broken link.
  • Nothing invented. Every line in the profile is something the user said, confirmed, or approved from a draft. Anything unknown becomes a TODO: line, never a plausible guess. Why: a guessed audience or a made-up disclaimer gets copied into hundreds of posts before anyone notices.
  • Compliance wording is theirs. If the business is regulated, the exact disclaimer text comes from the user, their regulator's published wording, or their compliance contact. You can suggest where to find it; you do not author it. Mark anything unconfirmed TODO: confirm with compliance.
  • Small batches. Three or four questions at a time, so it feels like a chat over coffee rather than a form. Offer multiple-choice options (an AskUserQuestion-style picker where the client supports it) with an "other" escape hatch.
  • Short is fine. A sharp two-line voice section beats a padded page. Write in the user's spelling locale.

The method

1. Open and offer a head start

Explain in two or three sentences what you're building and that it takes about ten to fifteen minutes (or five for the quick version). Then ask, in one batch:

  • Where should the workspace live? (default marketing/)
  • Full setup or quick setup? Quick covers Identity, Audience, Offer & goal, Voice, Platforms and the regulated yes/no; the rest become TODOs to fill later.
  • Is there existing material to draft from? Website URL, About page, LinkedIn or Instagram bio, a few past posts they're proud of, a brand guidelines PDF, a pitch deck.

If they give material, read it (fetch the URL, read the files or pasted text) and pre-draft as many sections as you can. Present drafts as "here's what I picked up from your site, is this right?" rather than asking questions you could have answered yourself. Label drafted lines so the user knows what to check.

2. Run the interview in batches

Work through references/interview-guide.md, one batch per turn. It has the questions, the multiple-choice options and the follow-ups for each schema section. Skip any question already answered by the material from step 1, and confirm the drafted answer instead.

Keep a running draft as you go. After each batch, reflect back the one or two lines you've written for that section so the user can correct it while it's fresh.

Things worth slowing down for:

  • Voice. Ask for two or three real posts or messages they wrote themselves and like. Pull the traits from those, not from adjectives they pick off a list. Adjectives describe the voice they want; samples show the voice they have.
  • I or we, per channel. People connect with a person, not a committee. A "we" post on a personal profile reads like a press release; an "I" post on a joint brand page leaves the partner out. Record the person per platform.
  • Boundaries. Ask what they never want to be seen talking about, and what claims they never make. Side projects and second businesses often need a firm line so one brand doesn't bleed into another.
  • Compliance. Ask whether the business is regulated. If yes or unsure, open the regulated-industries guidance in the writing skill (references/regulated-industries-example.md) to frame the questions about tiers. Ask about affiliate links and testimonials even for unregulated businesses, because advertising rules apply to everyone.
  • Likeness. If the user wants to appear in generated images, ask for three to six real photos (clear face, natural light, current look, at least one in colour) and copy them into brand/reference-images/. These are for likeness reference only and are never posted directly. Ask about consent for anyone else pictured.

3. Write the workspace

Create this structure (skip folders that exist):

marketing/
  brand/brand-profile.md
  brand/voice-learnings.md
  brand/performance-learnings.md
  brand/sources.md
  brand/reference-images/
  content/
  baselines/
  data/
  • brand-profile.md: from the template, filled with the confirmed answers, template comments removed, TODO: lines kept for gaps. Add a Last updated: YYYY-MM-DD line under the title.
  • voice-learnings.md, performance-learnings.md, sources.md: empty but structured, using each owning skill's template (voice-loop, platform-data, research) so the headings match what those skills expect. references/workspace-files.md says which template to use and has fallbacks if those skills aren't installed. Seed voice-learnings.md with the voice samples the user gave (quoted verbatim, dated) because those are the first real evidence of how they write.
  • reference-images/: the photos, plus a short README.md listing each file, what it shows and any rules (for example "colour photo, best for hair and skin tone").

4. Read it back

Show the user a compact summary: one line per section, plus the list of TODOs. Ask them to skim the profile itself. Offer the obvious next step (usually "want me to write your first post with the writing skill?").

Re-run mode (updating one section)

When a profile already exists:

  1. Read it and list the twelve sections with a one-line status each (filled, thin, or TODO).
  2. Ask which section to update, or offer to work through the TODOs.
  3. Run only that section's batch from the interview guide, showing the current text first so the user edits rather than starts again.
  4. Replace only that section. Leave every other section byte for byte as it was. Update the Last updated line.
  5. If the change affects other skills' behaviour (a new platform, a new disclaimer tier, a dropped boundary), say so in one line, for example "new platform added, so scheduling will need its posting route".

If the profile is missing headings or uses old names, offer to migrate it to the current schema, keeping all existing content under the right headings.

When another skill finds no workspace

Other skills say so in one line and offer to run this. If the user wants to carry on without it, those skills ask only the essentials (business, audience, platform, voice in a sentence, regulated or not) and suggest running brand-setup afterwards so the next piece doesn't repeat the questions.

Checks before it ships

Before handing over, check:

  • All twelve headings present, exact spelling, in order.
  • No invented facts: every line traces to something the user said, confirmed or approved.
  • Compliance wording is either confirmed or marked TODO: confirm with compliance.
  • Platforms table complete for each listed platform (handle, voice, route, links allowed).
  • Content pillar percentages add up to roughly 100.
  • Reference images copied and listed, consent noted for anyone else in them.
  • No em dashes, and the user's spelling locale throughout.

For a profile that will drive regulated content, run it through the review-loop skill before the first post goes out.

Adapting it

  • Agencies or multi-brand users: one workspace per brand. Keep brands apart; a shared profile is how one brand's claims end up on another's posts.
  • Couples and partnerships: record both people in Identity and set I/we per channel in Platforms.
  • Very small or brand-new businesses: quick setup, then revisit Content pillars and Visual identity after a month of posting, when platform-data has something to say.
  • Translating: set the spelling locale and language in Voice; every skill writes in it.

Learn more

A longer walkthrough, with examples for different kinds of business, lives at https://builtnotboosted.com/skills/brand-setup

Individual skills in this repo

This repo contains 6 individual skills — each has its own dedicated page.

builtnotboosted/image-generation

Produces finished social images with any image model (Google Gemini image, "Nano Banana", as the worked example) - feed statics, carousel slides, short-form video covers, YouTube and article thumbnails, local listing cards and animated statics. Bakes the headline into the scene instead of pasting text on top, keeps text inside the profile-grid safe zone, locks the likeness of real people to their reference photos, hits exact platform sizes, writes alt text and re-judges every re-roll on the filed file. Use when someone says "make an image for this post", "create a thumbnail", "make a cover from my video", "design a carousel", "animate this static", "the image model won't do 4:5", "my headline is getting cut off in the grid", or when a content pack needs its visuals.

builtnotboosted/research

Finds and checks topics, stories, facts and figures for content, then hands over a sourced research brief that the writing and video-scripting skills build from. Use when someone asks what to post about, wants ideas or angles, asks to research a topic, check a fact or statistic, find out what their audience is searching for or asking, react to news or a data release in their industry, plan around a calendar of announcements, or build or tidy their list of trusted sources. Trigger phrases include "research this", "what should I post this week", "find me a story", "is this stat right", "what are people searching for", "react to today's news", "content ideas for", "fact-check this post" and "update my sources". Every claim it returns carries a named source and a date, and it never invents a figure, quote, review or story.

builtnotboosted/review-loop

The quality gate every piece of content passes before it is filed, scheduled or posted. Runs a self-critique, then an independent review by a fresh subagent that sees only the finished work and the criteria (never the builder's reasoning), then fix and re-judge until there is a clean pass. Covers what to do when a piece can't pass unattended (hold, flag, ship the rest). Use when someone says "review this", "check this before it goes out", "QA the pack", "is this ready to post", "get a second pair of eyes on this", "proofread the captions", or when any other Built Not Boosted skill hands work over for checks.

builtnotboosted/scheduling

Gets finished posts into a social media scheduler (or posted by hand) safely. Queue-only with explicit timezone offsets, live dedup against scheduled and already-published posts, media readiness checks, correct post types (shorts, reels, carousels, first comments, manual-posting reminders), alt text, a link gate that sends copy back for a rewrite rather than stripping it, and confirmation before anything goes live. Writes and keeps a scheduling manifest for each pack. Works with any scheduler; Publer is the worked example. Use when someone says "schedule this", "queue these posts", "put this in Publer", "book in this week's content", "post this on Tuesday at 7", "move that post", "what's already scheduled", or hands over a finished content pack to go out.

builtnotboosted/video-editing

Edits filmed talking-head footage into release-ready vertical shorts (1080x1920) and 16:9 YouTube long-forms with ffmpeg and Whisper. Keeps exactly one clean read of every script line (no stumbles, restarts or double reads), never leaves a flat single shot (punch-ins, reframes, B-roll, motion graphics), captions from the finished audio, normalises loudness to -14 LUFS, makes a real-frame cover and runs a QA gate that re-transcribes the render and refuses delivery until it passes. Syncs extra cameras by sound and writes YouTube chapters. Use when someone says "edit my video", "cut this clip", "remove my retakes", "edit out the fluffs", "add captions to my video", "make a short from this footage", "sync my two cameras", "make a thumbnail from my video", or hands over raw footage and a script.

builtnotboosted/video-scripting

Writes video scripts that sound like the person who will say them, for shorts (Reels, TikTok, YouTube Shorts) and long-form YouTube. Warm hook rather than a shout, who-am-I and follow early, a retention shape that delays the answer, a soft earned close, length that follows the story, real-problem substance, reactive scripts for news and data releases. Output is teleprompter-ready (think-breaks, shot reminders) with short edit notes for the video-editing skill. Use when someone says "write me a script", "script a reel about", "I need a short on", "plan a YouTube video", "turn this post into a video", "react to today's news on video", "what should I say on camera", or before a filming session.

Skills relacionados