What is multimodels-mcp AGENTS.md?

A set of instructions for using the multimodels-mcp server with Codex. The server lets Codex delegate tasks to other AI models, including cloud providers and models running locally through LM Studio.

Use it to send tasks from Codex to Codex, DeepSeek, z.ai, OpenRouter, or local models. It describes the local Mac setup, commands, and configuration files.

It avoids handling each model through a separate workflow and provides a local panel for API keys and model selection. Results are returned to Codex for evaluation.

multimodels-mcp AGENTS.md — Instructions file for Codex and OpenCode by dpmadsen: A set of instructions for using the multimodels-mcp server

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods

npx agentmods add instructions/dpmadsen/multimodels-mcp/agents-md

Clone the repo

git clone --depth 1 https://github.com/dpmadsen/multimodels-mcp

Made for: Codex, OpenCode.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

README.md

[![agentmods](https://agentmods.dev/badge/instructions/dpmadsen/multimodels-mcp/agents-md.svg)](https://agentmods.dev/instructions/dpmadsen/multimodels-mcp/agents-md)

Your own site

<a href="https://agentmods.dev/instructions/dpmadsen/multimodels-mcp/agents-md"><img src="https://agentmods.dev/badge/instructions/dpmadsen/multimodels-mcp/agents-md.svg" alt="Measured on agentmods" height="20"></a>

Per session 1,237 This file is loaded in full into every session.

When invoked 1,237 The same file — it is already loaded in full.

Security scan A 0 findings. Scan, not verified.

Origin original No closer match found in the catalogue.

Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

Model	Per session	Once invoked
Fable 5	$0.01237	$0.01237
Opus 5	$0.00619	$0.00619
Sonnet 5	$0.00247	$0.00247
Haiku 4.5	$0.00124	$0.00124

Measured 5d ago against content hash e774b963d626, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

multimodels-mcp AGENTS.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 5d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

Origin

Copies of this mod

1 near-identical copy found in the catalogue:

multimodels-mcp CLAUDE.md — 92% identical, 11 lines differ

AGENTS.md · 61 lines

How it starts

The opening of the file, as written. The whole thing — 61 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Projeto: multimodels-mcp

English note: AGENTS.en.md

O que é

Servidor MCP que funciona como um "garçom" entre o Codex e outros modelos de IA: o Codex delega uma tarefa, o servidor processa com o modelo escolhido e devolve o resultado pro Codex avaliar. Modelos disponíveis: Codex (via assinatura do ChatGPT, pelo CLI codex já instalado no Mac), DeepSeek, z.ai e OpenRouter (via chaves de API) e modelos locais via LM Studio. Inclui um painel de controle local pra gerenciar chaves e habilitar/desabilitar modelos.

Pra quem é

Pro Daniel, uso pessoal, rodando só no Mac dele dentro do Codex.

Stack e comandos

Servidor: TypeScript + SDK oficial de MCP da Anthropic (transporte stdio, roda local)
Conector universal: padrão OpenAI-compatible (atende DeepSeek, z.ai, OpenRouter e LM Studio — provedor novo = 1 linha de configuração + chave no .env)
Codex: integração via CLI codex (usa a assinatura, sem custo de API)
Painel de configuração: shadcn/ui + Tailwind (Vite, pasta ui/), página local no navegador — abrir com npm run panel (http://127.0.0.1:4747); depois de mudar código da interface, npm run build:ui
Configuração de modelos: arquivo local config/models.json (sem banco de dados)
Chaves: SEMPRE no .env (nunca no models.json, nunca no código)
Rodar os testes: npm test (compila e testa; os testes vivem em src/*.test.ts)
Compilar depois de mudar código: npm run build (o Codex executa o resultado em dist/)
O servidor está registrado no Codex com o nome multimodels (escopo user, vale pra todos os projetos); conferir com Codex mcp list
Ferramentas expostas: list_models (cardápio) e delegate_task (delegação)

Como trabalhar comigo

Sobre mim

Eu NÃO sei programar. Já fiz projetos de vibecoding, mas preciso de direcionamento.
Explique tudo em português, com analogias do dia a dia. NUNCA assuma que sei termos técnicos.
Respostas objetivas. Se eu quiser mais detalhes, eu peço.

Como se comunicar comigo

Ao fazer algo, mostre: (1) o que fez, (2) como eu vejo/testo, (3) o que fazer se der erro.
Se der erro, explique primeiro em linguagem simples, DEPOIS sugira a correção.
Pedido vago? Faça perguntas até entender. Decisões importantes: apresente opções e explique as diferenças em termos simples.
A menos que eu diga "faça isso agora", trate meu pedido como conversa: pense comigo, proponha uma abordagem e só modifique arquivos depois que eu confirmar.

Read the full file on GitHub · 61 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

5d ago First seen · 61 lines · 1,237 tokens per session scan A e774b963d626

Subscribe to this mod's changes

multimodels-mcp AGENTS.md is an instructions file published in the GitHub repository dpmadsen/multimodels-mcp (19 stars, last pushed 1mo ago), licensed MIT. It adds 1,237 tokens to every session, about $0.0062 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.

Building agents

vscode buildNext.instructions.md

Working notes and architecture documentation for the new esbuild-based build system in build/next. Use when making changes to the new build pipeline (transpile/bundle commands, NLS plugin, source-map handling, resource copying, or self-hosting watch tasks).

microsoft/vscode · 6,785 tokens

multimodels-mcp AGENTS.md