skills.rest
michael-conrad
michael-conradCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

multimodal-dispatch

Probe Ollama model capabilities and dispatch tasks by content modality.

Updated Apr 24, 2026
Search Tags:#routing#multimodal#dispatch#ollama#sub-agent#model-probing#capability-snapshot
One-click install
npx skills add https://github.com/michael-conrad/.opencode --skill multimodal-dispatch
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: multimodal-dispatch
Source: https://github.com/michael-conrad/.opencode/tree/main/skills/multimodal-dispatch
Command: npx skills add https://github.com/michael-conrad/.opencode --skill multimodal-dispatch

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill provides a modality-aware routing framework that routes tasks to the most capable Ollama model based on content modality, using a cloud-first preference for text and vision models and coordinating sub-agent dispatch.

Core Features & Use Cases

  • Probes model capabilities and builds a capability snapshot for routing decisions.
  • Resolves content modalities against hints and dispatches per-modality tasks to the appropriate models.
  • Supports single-modality dispatch and multi-modality coordination via dispatch-multi with deterministic results.
  • Works end-to-end with probe, resolve, and dispatch tasks to enable reliable agent orchestration in OpenCode.

Quick Start

Route an incoming task by modality by probing capabilities, resolving the modality, and dispatching to the best model.

Frequently Asked Questions about multimodal-dispatch

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I route tasks to different models based on content modality?▼

Modality-aware routing resolves content modalities against hints and dispatches tasks to the most capable Ollama model. It uses a cloud-first policy for text and vision models to ensure deterministic results.

How does model capability probing work for task dispatch?▼

Model capability probing queries available Ollama models and builds a capability snapshot. This snapshot drives routing decisions by matching resolved content modalities to the most capable local or cloud model.

Can I coordinate multiple modalities in a single dispatch?▼

Yes, multi-modality coordination is supported via dispatch-multi. It dispatches sub-agents with modality-aware prompts and returns structured TaskResults for reliable downstream agent orchestration.

Does multimodal routing work with Ollama and OpenCode?▼

Yes, the routing framework integrates with Ollama for model probing and dispatches sub-agents within OpenCode. It requires no external dependencies to execute modality-aware task routing.

What is the best way to enforce deterministic results when dispatching sub-agents?▼

Deterministic execution is enforced by resolving content modalities, applying a cloud-first preference, and recording overrides. The framework returns structured TaskResults to guarantee reliable sub-agent outputs.

Why does modality-aware routing record overrides during task dispatch?▼

Recording overrides logs deviations from the cloud-first policy or selected model per modality. This ensures safe, deterministic task execution and provides an audit trail for downstream verification or research skills.

Related Skills

Explore More Software Engineering Skills→
tyronerosstyroneross
1

model-tiering

Select model tiers for subagent roles in a multi-phase build-loop.

Community
Intermediate
CommonCapitalCommonCapital
2

web-research

Plan targeted queries, gather sources, and generate structured reports with confidence indicators.

Community
Advanced
frankxaifrankxai
6

orchestration/parallel-execution

Coordinate parallel task execution across agents and tools with thread-safe dispatch and failure recovery.

Community
Intermediate
AesopScottAesopScott

selection-rubric

Implement a capability-based rubric for model selection and task routing.

Community
Advanced