Creator Teardown — Development skill for Claude Code
Claude Code / Cowork skill that tears down a creator's whole account: last 100-300 posts, outliers against their own normal, every post and sentence labelled by Jev, a one-page playbook, five rebuilt.
How to install Creator Teardown
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open ai-orchestrators/creator-teardown and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What Creator Teardown does
Claude Code / Cowork skill that tears down a creator's whole account: last 100-300 posts, outliers against their own normal, every post and sentence labelled by Jev, a one-page playbook, five rebuilt drafts and a checked review panel. Ships with zero identity: a first-use wizard builds your setup locally.
Alternatives in Development
- Open Claude Code — Nightly Claude Code CLI Decompile — Reverse Engineered & Rebuilt 210 ★
- Claude Code Leaderboard — This CLI automatically monitors your token usage and posts your stats to the leaderboard after each Claude Cod 191 ★
- Claude Code Rebuilt — A rebuilt, fully functional version of Anthropic's Claude Code CLI 120 ★
README
creator-teardown
A Claude Code / Claude Cowork skill that tears down a creator's whole account (their last 100-300 Instagram reels, TikToks, YouTube videos and Shorts, or LinkedIn posts) and finds what actually drives their numbers. Then it writes five versions for your brand.
Why this exists
Watching a creator's best posts tells you what they did. It does not tell you what worked. Their best posts share traits with their worst ones, they repost winners, and their account grows under them, so a hook style that looks like a formula is often a coincidence.
This skill answers with numbers instead of vibes:
- Measures every post against the creator's own normal at the time, so account growth never fakes an outlier
- Labels every post and every sentence the same way (hook type, structure, evidence, emotion, ending, the job each sentence does), so patterns are counted, not remembered
- Shows the miss rate, not just the hits: "9 of 22 broke out, 13 did not" is a test, not a formula
- Counts ideas, not posts, so a winner posted three times cannot fake a pattern
- Checks every pattern outside the winning shape, so a trait riding on the real cause gets folded into it
- Never invents a fact about you: every claim in a rebuild is proved by a file you named
Who a run is for
Every run starts with one question: who is this for?
| Answer | What you get |
|---|---|
| You (your own brand) | The full report, plus five rebuilds for whichever of your brand profiles you pick |
| A client | The skill builds the client's brand profile first, from their own files, posts and your notes, then writes the five rebuilds for them. One client per run, and every draft comes to you before anything reaches them |
| The creator | A report written to the creator about their own account: "Your best posts", "Keep doing this", "Do less of this". No rebuilds |
What you get
One self-contained HTML report with seven tabs, in dark and light themes, that
Related Skills
Claude Sprint Agent
A wrapper script that runs an unattended Claude Code "sprint" in headless mode, monitors the output for a DONE
Environment Wrapup
Tears down the current agent environment and runs /session-wrapup.
Writing Hooks
Claude Code / Cowork skill that generates scroll-stopping opening lines for any content. 5 hooks per run acros
Writing LinkedIn Posts
Claude Code / Cowork skill that drafts a complete LinkedIn post: scored hook, body on a deliberate rhythm, CTA
Gemini Super Gem Creator
Claude Cowork skill for building expert-level Gemini Gems with persistent Drive memory and NotebookLM groundin
Friction
Read the whole session and find what slowed the work down
Related Agents
Coworker
Run one cowork.py operation against Codex in this checkout and return its output unchanged.
Cowork Comparison Lead
Kortix is open source. It is the open-source AI Management System and the leading open-source alternative to C
Review Test Structure
Reviews test code for isolation, independence, readability, documentation quality, and proper setup/teardown.