agentleFS
mehbylabsmehbylabs/creator-skills/skills/creator-setup/SKILL.md

creator-setup

Skill7 starsChanged 16 days agoMITView on GitHub

Detect which media tools this user can use (Higgsfield CLI, Gemini/Google API key, OpenAI API key, ElevenLabs API key, Antigravity CLI agy, Codex CLI, Gemini web app, Google Flow) and save their creator-skills preferences: quality vs cost profile, provider priority per task, disabled providers, default voices, brand defaults. Guides connecting API keys (hidden prompt, verified, stored in a private .env) and OAuth logins. Use when the user says "set up creator skills", "connect my API key", "add my Gemini/OpenAI/ElevenLabs/Azure key", "log in to Higgsfield/Codex", "what tools do I have", "change provider order", "use my subscription first", "enable browser use", or a generation skill reported missing or invalid providers. Not for generating media.

  • Reads credentials

The file

We split this skill into the rules it gives an agent

6 rules. About 0 are common, 1 is distinctive, and 5 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 5

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Don't print JSON or key fragments beyond the masked last 4.

    line 34
  2. Very specific:

    Ask which missing providers they want (suggest the cheapest path to cover image, video, and voice).

    line 38
  3. Very specific:

    When they're done, run python3 scripts/connect.py verify <provider> and report.

    line 50
  4. Very specific:

    Never: ask for keys in chat, read .env files, or put keys on a command line.

    line 52
  5. Very specific:

    Write ~/.config/creator-skills/config.json (or ./creator-skills.json for one project) following references/config.md.

    line 66
  6. Very specific:

    If detect reports a broken tool config (e.g.

    line 79

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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

Creator Setup

Generation skills work with no config (auto-detect + quality profile). This skill makes the user's choices explicit and fixes what's missing.

Step 1 — Detect and verify

python3 scripts/creator.py detect
python3 scripts/connect.py verify

Show a short table: provider · status (connected / missing / invalid) · what it unlocks. Don't print JSON or key fragments beyond the masked last 4.

Step 2 — Connect what the user wants

Ask which missing providers they want (suggest the cheapest path to cover image, video, and voice). For each, follow references/connect.md:

  • Give the exact command with this skill's absolute script path for the user to run in their own terminal (python3 <abs>/scripts/connect.py google-api). It prompts for the key hidden, verifies it, saves to ~/.config/creator-skills/.env (chmod 600).
  • OAuth/CLI providers (higgsfield, codex, agy, google-cloud-tts) — same command; it runs their browser login.
  • Browser apps (gemini-browser, flow-browser, aistudio-browser) are on by default as the last fallback; the user only signs in once in the agent's browser. If this agent has no browser / computer-use tool, or the user doesn't want computer use, set "browser": {"enabled": false} (or narrow browser.apps) in the config.
  • When they're done, run python3 scripts/connect.py verify <provider> and report.

Never ask for keys in chat, read .env files, or put keys on a command line. If a key is pasted into chat, tell the user to rotate it and use the terminal flow.

Step 3 — Choose (one question at a time, only what's unclear)

  1. Profile — quality (APIs/CLIs first) or cost (subscription tools — agy, codex, Flow — before pay-per-call APIs). Explain with references/providers.md.
  2. Order — only if the user states a preference ("always agy first for images").
  3. Disable — providers they never want used or billed.
  4. Voice — optional default voice for voiceovers (references/audio.md).
  5. Brand — optional colors, heading font file, logo path for thumbnails/ads.

Step 4 — Write and confirm

Write ~/.config/creator-skills/config.json (or ./creator-skills.json for one project) following references/config.md. Never write secrets into it. Then:

for t in image video speech music sfx; do python3 scripts/creator.py plan --task $t; done

Report one line per task: the order that will be used, or "no provider — <cheapest fix>".

Rules

  • Secrets only flow through connect.py in the user's terminal, a password-manager pipe (--stdin), or env vars the user sets themselves.
  • If detect reports a broken tool config (e.g. Codex config.toml parse error), tell the user the file; don't edit other tools' configs without permission.

References

  • references/connect.md — connecting keys, OAuth logins, browser apps; security rules
  • references/providers.md — matrix, default orders, fallback rules
  • references/config.md — config schema
  • references/audio.md — voices

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.