What is multi-model-companion CLAUDE.md?

A set of Claude Code instructions for choosing among several AI models according to the task. It recommends a low-cost local model for suitable work and paid models for escalation.

It is for configuring model choices for code review, planned implementation, research summaries, tests, bulk work, parallel opinions, and difficult decisions.

It provides routing rules so routine work, difficult code, research, reviews, and tests are sent to an appropriate model.

How Reederey87/multi-model-companion configures Claude Code

Reuse

Borrowing it

Nothing to install: this file belongs to Reederey87/multi-model-companion. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file

curl -O https://raw.githubusercontent.com/Reederey87/multi-model-companion/main/CLAUDE.md

Clone the repo

git clone --depth 1 https://github.com/Reederey87/multi-model-companion

Made for: Claude Code.

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/reederey87/multi-model-companion/claude-md/github.svg)](https://agentmods.dev/instructions/reederey87/multi-model-companion/claude-md)

Your own site

<a href="https://agentmods.dev/instructions/reederey87/multi-model-companion/claude-md"><img src="https://agentmods.dev/badge/instructions/reederey87/multi-model-companion/claude-md/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

Your own site · 80×15

<a href="https://agentmods.dev/instructions/reederey87/multi-model-companion/claude-md"><img src="https://agentmods.dev/badge/instructions/reederey87/multi-model-companion/claude-md.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 633 This file is loaded in full into every session.

When invoked 633 The same file — it is already loaded in full.

Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.

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 5.1	$0.00633	$0.00633
Opus 5.5	$0.00253	$0.00253
Sonnet 5.5	$0.00127	$0.00127
Haiku 4.5	$0.00063	$0.00063

Measured 2mo ago against content hash fa45e22630a8, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

multi-model-companion CLAUDE.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 2mo 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.

CLAUDE.md · 49 lines

How it starts

The opening of the file, as written. The whole thing — 49 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Model picking (multi-model-companion)

Drop this file into a project root (or merge into ~/.claude/CLAUDE.md) so Claude Code prefers the companion tools correctly.

Default: use the local / cheap provider (deepseek slot — whatever you mapped in DEEPSEEK_MODEL) for anything that fits the context you supply. Paid subscription models (Codex, Grok) are escalations, not defaults.

Routes by use case

Use case	Route
Code review	`delegate_review(deepseek)` first; add `codex` (tricky logic) or `grok` (adversarial) for high-risk diffs; anything that ships gets +1 native Claude pass
Implement from a written plan	`delegate_patch(deepseek)`; `codex` if the spec is fuzzy or the first pass misses
Research synthesis	Claude gathers (web / files), then `delegate_analysis(deepseek)`; `grok` to challenge, `codex` for accuracy
Tests / bulk / repetitive	`deepseek` via `delegate_patch` or `compare_models(mode="tests")`
Parallel opinions	`compare_models` — always include `deepseek`; add paid providers only when opinions materially matter
User-facing work (UI, copy, API design)	native Claude — not delegated
Hardest judgment	native Claude (strongest available) — one pass per task, never the default worker

Escalation

Local/cheap first, always. Escalate only on evidence: thin/wrong output, fuzzy spec, high blast radius.
codex: strong agentic executor for precise patches and hard debugging.
grok: good paid cross-check / adversarial review.
Confirm live model IDs with list_companion_models — names churn.

Rules

Claude coordinates: decomposes, picks the model, selects context, verifies, applies, tests.
Companion output is a proposal — review every diff line-by-line, then run format/tests/checks. Never trust the narrative.
Bounded context only; never secrets, credentials, .env, key material.
Companions cannot read files, run shell, or use your other MCP tools — gather first, then delegate the digest.
Run health_check before delegation-heavy sessions.
Thinking models: keep max_output_tokens ≥ 1024 (reasoning spends budget first). Defaults when omitted: deepseek=max, grok=max, codex=high.
One model for ordinary delegation; compare_models only when independent opinions materially improve the result.
A result with content_source: "reasoning" is leaked chain-of-thought, not an answer — retry with a larger max_output_tokens; never treat it as a deliverable.

Read the full file on GitHub · 49 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.

2mo ago First seen · 49 lines · 633 tokens per session scan A fa45e22630a8

Subscribe to this mod's changes

multi-model-companion CLAUDE.md is an instructions file published in the GitHub repository Reederey87/multi-model-companion (0 stars, last pushed 3mo ago), licensed MIT. It adds 633 tokens to every session, about $0.0025 per session on Opus 5.5. 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 2026-08-31.

Building agents

multi-model-companion CLAUDE.md next to 6 instructions from other repositories
Mod	Type	Stars	Last push	Licence	Scan	Tokens	Works with
multi-model-companion CLAUDE.mdReederey87/multi-model-companion	Instructions file	0	3mo ago	MIT	A	633	Claude Code
vscode buildNext.instructions.mdmicrosoft/vscode	Instructions file	195k	today	MIT	A	9,719	GitHub Copilot
vscode oss-third-party-notices.instructions.mdmicrosoft/vscode	Instructions file	195k	today	MIT	A	5,516	GitHub Copilot
next.js AGENTS.mdvercel/next.js	Instructions file	143k	today	MIT	A	7,461	Codex, OpenCode
langchain AGENTS.mdlangchain-ai/langchain	Instructions file	149k	today	MIT	A	4,680	Codex, OpenCode
vscode copilot-instructions.mdmicrosoft/vscode	Instructions file	195k	today	MIT	A	3,093	GitHub Copilot
vscode model-prompts.instructions.mdmicrosoft/vscode	Instructions file	195k	today	MIT	A	2,022	GitHub Copilot

multi-model-companion: Instructions file for Claude Code