multimodel
Multi-model collaboration and orchestration. Runs a task across several AI models in parallel and aggregates their blind APPROVE/REJECT votes, or delegates it to one model running a full Claude Code session.
| Version | 3.7.1 |
| Marketplace | magus |
| Commands | 2 |
| Subagents | 1 |
| Skills | 17 |
| MCP server | no |
| Hooks | yes |
Install
claudeup
Find multimodel on the Plugins tab and turn it on. claudeup registers the magus
marketplace if you do not have it, and installs what the plugin needs to actually run —
binaries, MCP servers, CLI tools.
For a team, press s to save your plugins as a profile and commit it. Everyone else runs
claudeup install — see Teams and profiles.
Prefer to do it by hand? Installing Magus has the manual path.
How to run these: second opinions and delegation — what to type, what it asks you, and which options change the outcome.
When to reach for it
- Use when launching multiple agents, reading many files, or removing sequential execution bottlenecks —
batching-patterns - Use when the user mentions claudish, OpenRouter, or external AI models —
claudish-usage - Use when routing work to an external model —
delegate-patterns - Use when dealing with external model timeouts, API failures, partial success, user cancellation, or graceful degradation —
error-recovery - Use when orchestrating 3+ agents or multi-phase work —
hierarchical-coordinator - Use when orchestrating workflows that generate multiple files (designs, reviews, reports) to prevent file collisions across concurrent or sequential sessions with unique session directories —
session-isolation
Commands
| Command | What it does |
|---|---|
/multimodel:delegate | | |
/multimodel:team | | |
Subagents
Dispatched with the Agent tool, each in its own context window.
| Agent | What it does |
|---|---|
multimodel:deep-analyst | Orchestrates a multi-source deep investigation — parallel web research, local code and data evidence, and optional independent passes by subagents or external models — then consolidates ever… |
Skills
| How you get it | |
|---|---|
| ● | Claude can reach for it on its own |
| ○ | You invoke it by name |
Why a skill lands in one row or another
| Skill | What it covers | |
|---|---|---|
| ● | multimodel:batching-patterns | Batches related operations into single messages for maximum parallelism. Use when launching multiple agents, reading many files, or removing sequential execution bottlenecks. |
| ● | multimodel:claudish-usage | Routes Claudish between MCP tools (team, create_session for /team and /delegate) and the CLI for direct usage. Use when the user mentions claudish, OpenRouter, or external AI models. |
| ● | multimodel:error-recovery | Handle errors, timeouts, and failures in multi-agent workflows. Use when dealing with external model timeouts, API failures, partial success, user cancellation, or graceful degradation. |
| ● | multimodel:multi-agent-coordination | Coordinate multiple agents in parallel or sequential workflows. Use when running agents simultaneously, delegating to sub-agents, switching between specialized agents, or managing agent sele… |
| ● | multimodel:multi-model-validation | Runs the same task across multiple AI models in parallel and aggregates verdicts. Use when the user wants a second opinion, multi-expert validation, or consensus from Grok, Gemini, GPT-5, or… |
| ● | multimodel:task-complexity-router | Routes tasks to model tiers (haiku/sonnet/opus) by complexity. Use for model selection, complexity routing, picking a tier, or optimizing API cost. |
| ● | multimodel:task-orchestration | Track progress in multi-phase workflows with Tasks system. Use when orchestrating 5+ phase commands, managing iteration loops, tracking parallel tasks, or providing real-time progress visibi… |
| ○ | multimodel:agent-enforcement | | |
| ○ | multimodel:delegate-patterns | Delegating plugin commands and tasks to external models via claudish — capability discovery, pre-flight question gathering, and prompt assembly. Use when routing work to an external model. |
| ○ | multimodel:hierarchical-coordinator | Prevents goal drift in long multi-agent runs by validating outputs against the original objective at checkpoints. Use when orchestrating 3+ agents or multi-phase work. |
| ○ | multimodel:hooks-system | Lifecycle hook patterns — PreToolUse, PostToolUse, UserPromptSubmit, Stop, SubagentStop. |
| ○ | multimodel:model-tracking-protocol | MANDATORY tracking protocol for multi-model validation. Creates structured tracking tables BEFORE launching models, tracks progress during execution, and ensures complete results presentatio… |
| ○ | multimodel:performance-tracking | Track agent, skill, and model performance metrics for optimization. Use when measuring agent success rates, tracking model latency, analyzing routing effectiveness, or optimizing cost-per-ta… |
| ○ | multimodel:proxy-mode-reference | Reference guide for using external AI models via claudish MCP tools and CLI. Orchestration workflows (/team, /delegate) use MCP tools. Direct usage uses CLI. |
| ○ | multimodel:quality-gates | Implement quality gates, user approval, iteration loops, and test-driven development. |
| ○ | multimodel:session-isolation | Use when orchestrating workflows that generate multiple files (designs, reviews, reports) to prevent file collisions across concurrent or sequential sessions with unique session directories. |
| ○ | multimodel:task-external-models | Quick reference for running external models in orchestration. They are invoked via claudish MCP tools (team, create_session), never the CLI. Use when unsure how to specify an external model. |
Hooks
This plugin installs hooks. They run automatically and are the usual first place to look if its behaviour stops firing.
Source
plugins/multimodel/ on GitHub, including its full README.