agentleFS
Sign inStart free
linxulelinxule/carrel/skills/model-teammates/SKILL.md

model-teammates

Skill2 starsChanged 6 days agoMITView on GitHub

This skill should be used when a researcher wants to add, remove, or review multi-model agent teammates (Codex/ChatGPT, Gemini, Kimi) inside Claude Code. Triggers on 'teammates', 'multi-agent', 'multi-model', 'second opinion', 'add Gemini', 'use ChatGPT here', 'orchestrate models', or when `/carrel-setup` enters Phase 5b.

  • Installs packages

The file

We split this skill into the rules it gives an agent

4 rules. About 0 are common, 1 is distinctive, and 3 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.

  • 3

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    If the researcher already pays for ChatGPT Plus/Pro or Gemini Advanced, the teammate runs on that subscription — no extra cost.

    line 18
  2. Very specific:

    Read .carrel/environment.json → model_teammates.

    line 22
  3. Very specific:

    Verify via the plugin's own setup command if it has one (/codex:setup, /kimi:setup.

    line 37
  4. Very specific:

    After each install / defer / decline / removal, update model_teammates in .carrel/environment.json and regenerate the dashboard:

    line 55

≈ “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

model-teammates

Guides researchers through adding foundation models as teammates inside Claude Code via three community plugins. Most researchers don't realize this is possible — the interview protocol's "About Model Teammates" beat carries the spoken script; this skill owns the operational details.

Research moves

TeammateReach for whenPlugin
Codex (ChatGPT)Adversarial / second-opinion review on a claim or methods paragraphopenai/codex-plugin-cc
GeminiLong-context / whole-corpus synthesis (hundreds of PDFs, large codebases)thepushkarp/cc-gemini-plugin
KimiDelegated bug hunts, refactors, free-form Q&A with CLI-auth privacylinxule/kimi-plugin-cc

If the researcher already pays for ChatGPT Plus/Pro or Gemini Advanced, the teammate runs on that subscription — no extra cost. Kimi's CLI handles auth locally.

Orientation

Read .carrel/environment.json → model_teammates. Status values:

  • configured — plugin installed + verified
  • interested — said yes but deferred install
  • skipped — declined; don't re-ask
  • (absent) — not yet discussed

Also read sensitivity on the profile. All three teammates route prompts through external providers (OpenAI, Google, Moonshot) — including Kimi despite its local CLI auth. HIGH sensitivity → don't propose; record skipped. MEDIUM → surface but confirm consent. LOW → proceed.

Install flow

For each teammate the researcher chooses:

  1. Install the provider's CLI and log in (researcher runs the commands — interactive auth needs human hands).
  2. Install the Claude Code plugin from the marketplace and reload.
  3. Verify via the plugin's own setup command if it has one (/codex:setup, /kimi:setup; Gemini has no dedicated setup — run a small query).

Plugin install commands inside Claude Code:

/plugin marketplace add openai/codex-plugin-cc        && /plugin install codex@openai-codex        && /reload-plugins && /codex:setup
/plugin marketplace add thepushkarp/cc-gemini-plugin  && /plugin install cc-gemini-plugin@cc-gemini-plugin && /reload-plugins
/plugin marketplace add linxule/kimi-plugin-cc        && /plugin install kimi@kimi-marketplace     && /kimi:setup

CLI install commands (each upstream README is authoritative — Claude fetches on demand if anything has drifted):

  • Codex: npm install -g @openai/codex → codex login
  • Gemini: npm install -g @google/gemini-cli (or brew install gemini-cli) → gemini auth
  • Kimi: npm install -g kimi-cli → kimi login (upstream: https://github.com/MoonshotAI/kimi-cli)

State writeback

After each install / defer / decline / removal, update model_teammates in .carrel/environment.json and regenerate the dashboard:

carrel vault dashboard --vault <path> --force

Removing a teammate

Uninstall the plugin (/plugin uninstall ...), optionally remove the CLI, delete the teammate's key from model_teammates in environment.json (or set to skipped if you want to remember "don't re-ask"), regenerate the dashboard.

Related

  • Command: /carrel-teammates (this skill's standalone entry)
  • Setup: /carrel-setup Phase 5b invokes this skill during onboarding
  • Spec: planning/specs/013-model-teammates.md

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.

Check your setup, free