kevinbadi/social-agents/brand-interview

Claude Code and agent skill by kevinbadi/social-agents

Trust tier 3
  • community
Source check pending

Checking that the SKILL.md file is still in the repository.

Install

npx vskill@latest install kevinbadi/social-agents/brand-interview

SKILL.md

brand-interview

The brand pack is the one file every caption, description, and CTA flows from. Onboarding does not collect it — you do, in chat, where you can follow up. Run this the moment social-agents/BRAND.md is missing, before anything else.

How to run it

One question at a time. Wait for the answer, reflect it back in a line, then move on. Never batch the questions into a form, never fill in guesses. If an answer is thin, ask one follow-up, then move on. Eight questions, about five minutes:

  1. About — "What is this brand actually about? What are you marketing?"
  2. Offers — "What do you sell, and where does each thing live?" One offer per line: link + what it is. No link yet is fine; note it as such. Every CTA you ever write points at one of these.
  3. Voice, three adjectives — "Your voice in three adjectives."
  4. Voice, one never — "One thing my copy should never sound like." (corporate, thirsty, salesy…)
  5. Emoji policy — none / sparingly (max one per caption) / free, part of the voice.
  6. Hashtag policy — none / a few relevant ones (2-4) / aggressive.
  7. Audience — "Target audience in one sentence."
  8. Competitors — up to five handles or URLs to watch. Empty is fine.

This workspace is ONE CreatorOS API key, one brand. If the human runs several brands (e.g. an agency), each has its own workspace: interview for THIS one only, in its own terms.

Write the file

Write social-agents/BRAND.md in exactly this shape, then read it back and ask for sign-off. Edit on request; the latest file always wins.

md
# Brand Pack
Social Agents read this before writing anything. Every caption, description, and
CTA flows from here. Edit freely — Social Agents always use the latest version.
## What this brand is about
<about>
## What we sell — products, services & CTA destinations
-<what it is> — <link>
-<what it is> _(no link yet)_
## Voice
-Sounds like: <adj>, <adj>, <adj>
-Never: <never>
-Emoji policy: <policy>
-Hashtag policy: <policy>
## Target audience
<audience>
## Competitors to watch
-<handle or URL>
Research findings live in `knowledge/COMPETITORS.md` — ask Social Agents to refresh them any time.

After sign-off

  • If competitors were named, research them (content mix, cadence, hooks, gaps) and write social-agents/knowledge/COMPETITORS.md.
  • Then the infrastructure question: where should automations live — local (this machine, awake at scheduled times) or a Railway worker you provision (provision-railway skill, from a Railway API token)? Save automationTarget and timezone to social-agents/social-agents.json. Local is a fine answer; never lecture on cost.
  • Then the automation menu, one item at a time, only what they approve (automations skill).

Judgment rules

  • No captions, no posts, no funnels before BRAND.md exists and is signed off.
  • Never invent an offer or a link. No link means no CTA to it.
  • Keep it a conversation: short questions, plain text, no forms.

Security

  • Security scanPassed, 95 of 100Pattern checks on the skill's files.
  • Model reviewPassed

Quality evaluation

  • Task evaluationNot runNo evaluation yet. Run one in Skill Studio.

Source

Popularity

Repository stars
55
Repository forks
19
7-day trend score
25

Stars and forks are for the whole GitHub repository, not this skill alone. Not refreshed yet.

Works with

Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.

  • Claude Code
  • Cursor
  • GitHub Copilot
  • Windsurf
  • Codex
  • Gemini CLI
  • Cline
  • Amp
  • Roo Code
  • Goose
  • Continue
  • Kilo Code
  • Junie
  • Trae
  • OpenHands
  • Qwen Code
  • AI Agent
  • Tabnine Chat
  • Sourcegraph Cody
  • Amazon Q
  • Firebase Genkit
  • Sweep
  • Devin
  • Phind
  • Mentat
  • Claude Desktop
  • Aider
  • GPT Pilot
  • Plandex
  • bolt.new
  • v0
  • Replit Agent
  • Anse
  • TypingMind
  • BetterChatGPT
  • LibreChat
  • Warp AI
  • Pieces
  • Fig AI
Embed badge
vskill VERIFIED badge
[![vskill VERIFIED](https://verified-skill.com/api/v1/skills/kevinbadi/social-agents/brand-interview/badge)](https://verified-skill.com/skills/kevinbadi/social-agents/brand-interview)

Something wrong with this skill or its page? Tell us.