skills.rest
tribgames
tribgamesOFFICIAL
Explore all skills by this author →
GitHub

local-provider

Install, inspect, repair, and remove Mixdog-managed local GGUF models.

6|Updated Jun 25, 2026
Search Tags:#llm#gpu#gguf#local models#hugging face#offline inference#model installation
One-click install
npx skills add https://github.com/tribgames/mixdog --skill local-provider-tribgames
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: local-provider
Source: https://github.com/tribgames/mixdog/tree/main/src/defaults/skills/local-provider
Command: npx skills add https://github.com/tribgames/mixdog --skill local-provider-tribgames

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Running local GGUF models involves hunting Hugging Face repositories, verifying licenses and revisions, checking GPU and disk fit, and recovering from interrupted downloads. This Skill manages that entire lifecycle through Mixdog's installer instead of manual shell downloads or configuration edits. ## Core Features & Use Cases - Model Discovery and Inspection: Search Hugging Face repositories, inspect exact GGUF files for size, license, pinned revision, SHA-256, context window, and VRAM estimates before committing to a download. - Managed Installation and Recovery: Start resumable background runtime and model downloads, monitor progress, cancel or resume paused jobs, and verify or repair model files with checksum validation. - Safe Lifecycle Operations: Activate the provider, adjust idle unload TTL, and delete models only after explicit confirmation with expiring tokens, without touching user files or bypassing failed checks. - Use Case: A user wants an offline coding model. The Skill checks GPU and disk status, inspects a candidate GGUF from Hugging Face, registers it after license approval, downloads the weights in the background, and routes the conversation to the local model once a real response is observed. ## Quick Start Ask the agent to check local provider status and find a GGUF model that fits your GPU, then install it through the managed installer.

Frequently Asked Questions about local-provider

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

FAQPage Schema
How do I install a local GGUF model in Mixdog?▼

First run setup with action status on the local-provider domain to check platform, GPU, and disk. Then search Hugging Face, inspect the exact GGUF file, register it with license acceptance, and start the model installation phase with the returned model id.

How do I find which GGUF model fits my GPU?▼

Inspect the chosen GGUF file with your intended context window to get a VRAM estimate based on the actual file header. The runtime reads architecture and context metadata and compares the estimate against detected free GPU memory before loading.

Can I resume an interrupted local model download?▼

Yes, downloads are resumable. A paused or failed job can be resumed with start_local_installation using its phase and model id, and after an app restart partial files appear as resumable installations without an active job id.

Does the local provider support vision or adjustable reasoning?▼

No, managed local inference is currently text-only and uses model-default reasoning settings. Tool support is reported as unverified until metadata or the server's capability endpoint confirms it, and unsupported attachments are rejected before transmission.

Why did my local model download fail with a checksum error?▼

A checksum failure means the downloaded file failed integrity validation and must not be bypassed. Use maintain_local_model with operation verify to check integrity, or operation repair for an approved redownload, which retains the previous file if it fails.

How do I safely delete a local model?▼

Read local_model_details to see exact file paths and size, obtain explicit approval, then call delete_local_model with the recent confirmation token. Deletion is permanent, not Recycle Bin based, and in-use models cannot be force deleted.

Related Skills

Explore More Software Engineering Skills→
HermesagentsHermesagents

hermes-agent-setup

Guide Hermes Agent setup through an interactive CLI configuration wizard.

Community
Intermediate
blueskies1818blueskies1818
1

llama-cpp

Locate and execute llama.cpp models for local inference.

Community
Intermediate
ericdahl-devericdahl-dev

benchmark-models

Compare Claude, GPT, and Gemini model latency, tokens, and quality.

Official
Advanced
freedomw1987freedomw1987
2

comfyui

Deploy, configure, and execute ComfyUI generative media workflows locally or in the cloud.

Community
Advanced