model-orchestrator

model-orchestrator is an agent for Claude Code from LeekJay/claude-skills-plugin. It costs 61 tokens per session (3,444 once invoked), scanned A, original, MIT.

An AI agent that analyzes a task, selects a suitable model, and prepares a prompt for that model. It is the decision layer in the smart model router.

In plain words
What is it for?
Use it to choose among Gemini CLI, GPT-5.2, Claude Opus, Claude Sonnet, and Claude Haiku for coding, frontend work, large-codebase analysis, architecture, security, or lightweight tasks.
Why use it?
It reduces the need to decide manually which model should handle each task. Its stated focus is making quick routing decisions based on model capabilities.

Agent for Claude Code

Written for Claude Code: shipped in a Claude Code plugin. Also seen: model in frontmatter; mentions subagents; mentions Codex.

Part of the smart-model-router plugin — 1 skill, 3 agents, 1 hook shipped together

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 agents/leekjay/claude-skills-plugin/model-orchestrator
Clone the repo
git clone --depth 1 https://github.com/LeekJay/claude-skills-plugin

Made for: Claude Code.

Or install smart-model-router, the plugin that ships this one along with the rest of its 1 skill, 3 agents, 1 hook.

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.

agentmods badge for model-orchestrator

README.md
[![agentmods](https://agentmods.dev/badge/agents/leekjay/claude-skills-plugin/model-orchestrator.svg)](https://agentmods.dev/agents/leekjay/claude-skills-plugin/model-orchestrator)
Your own site
<a href="https://agentmods.dev/agents/leekjay/claude-skills-plugin/model-orchestrator"><img src="https://agentmods.dev/badge/agents/leekjay/claude-skills-plugin/model-orchestrator.svg" alt="Measured on agentmods" height="20"></a>
Per session 61 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 3,444 The whole file, excluding the scripts and references it only reads on demand.
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.

ModelPer sessionOnce invoked
Fable 5.1 $0.00061 $0.03444
Opus 5 $0.00030 $0.01722
Sonnet 5 $0.00012 $0.00689
Haiku 4.5 $0.00006 $0.00344

Measured 5d ago against content hash ccf70ea4103d, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-06, from the pricing page.

Security

Grade A, and why

model-orchestrator 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.

plugins/smart-model-router/agents/model-orchestrator.md · 450 lines

How it starts

The opening of the file, as written. The whole thing — 450 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Model Orchestrator Agent

You are a lightweight decision-making agent powered by Haiku. Your role is to quickly analyze tasks and determine the optimal model for execution, while optimizing prompts for the target model.

Your Mission

Provide fast, accurate model routing decisions with optimized prompts. Minimize token usage while maximizing decision quality.

Core Principles

  1. Speed First - Make quick decisions, don't over-analyze
  2. Capability First - Route to the strongest model likely to succeed
  3. Quality Assurance - Ensure task will succeed with selected model
  4. Prompt Optimization - Tailor prompts for target model's strengths

Model Profiles

Available Models

Model Capability Speed Best For
Gemini CLI Top-tier (UI + huge context) Medium Frontend/UI, very large context, whole-repo understanding
Claude Opus 4.5 Highest reasoning Slow Architecture, security-critical, high-stakes decisions
GPT-5.2 Top-tier coding Fast Most coding/refactor/bugfix tasks (DEFAULT) ⭐
Claude Sonnet 4.5 Top-tier Claude features Medium Claude-specific workflows, artifacts
Claude Haiku 4.5 Good, lightweight Fastest Simple/realtime tasks, light reviews, sub-agents

Model Capabilities Matrix

Task Type GPT-5.2 Haiku Gemini Sonnet Opus
Code Review ★★★ ★★ ★★ ★★★ ★★★
Template Gen ★★★ ★★ ★★ ★★ ★★
Complex Coding ★★★ ★★ ★★ ★★ ★★★
Frontend UI ★★ ★★ ★★★ ★★ ★★
Large Context ★★ ★★★ ★★ ★★
Multi-file Refactor ★★★ ★★ ★★ ★★★ ★★★
Architecture ★★ ★★ ★★ ★★ ★★★
Security Analysis ★★ ★★ ★★ ★★ ★★★

Read the full file on GitHub · 450 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.

  1. 5d ago First seen · 450 lines · 0 tokens per session scan A ccf70ea4103d

Subscribe to this mod's changes

model-orchestrator is an agent published in the GitHub repository LeekJay/claude-skills-plugin (4 stars, last pushed 8mo ago), licensed MIT. It adds 61 tokens to every session and 3,444 once invoked, about $0.0003 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 2026-08-31.