CtrlK
BlogDocsLog inGet started
Tessl Logo

huggingface-local-models

Use to select models to run locally with llama.cpp and GGUF on CPU, Mac Metal, CUDA, or ROCm. Covers finding GGUFs, quant selection, running servers, exact GGUF file lookup, conversion, and OpenAI-compatible local serving.

90

1.25x
Quality

93%

Does it follow best practices?

Impact

73%

1.25x

Average score across 3 eval scenarios

SecuritybySnyk

Advisory

Suggest reviewing before use

SKILL.md
Quality
Evals
Security

Quality

Content

87%

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

This is a well-crafted skill that provides highly actionable, concise guidance for finding and running GGUF models with llama.cpp. Its strengths are excellent token efficiency, concrete executable examples throughout, and good progressive disclosure to reference files. The main weakness is the workflow section lacks explicit validation checkpoints — particularly around verifying downloads, confirming model compatibility, and validating conversion output before use.

Suggestions

Integrate the smoke test as an explicit validation step in the Default Workflow (e.g., after step 5: 'Verify: curl localhost:8080/v1/models to confirm the model loaded')

Add a validation checkpoint after the conversion step (step 7) such as verifying the GGUF file size or running a quick inference test before proceeding

DimensionReasoningScore

Conciseness

The content is lean and efficient throughout. It assumes Claude knows what GGUF, llama.cpp, and quantization are without explaining them. Every section provides actionable commands or URLs without padding. No unnecessary preamble or concept explanations.

3 / 3

Actionability

Fully executable commands throughout: install commands, search URLs, CLI invocations with real repo/quant examples, curl commands for smoke testing, and conversion pipelines. Everything is copy-paste ready with concrete repo names and flags.

3 / 3

Workflow Clarity

The Default Workflow section provides a clear 7-step sequence with logical ordering and fallback paths. However, there are no explicit validation checkpoints — no step to verify the download succeeded, confirm the model loaded correctly, or validate conversion output before proceeding. The smoke test exists but is separate from the workflow, not integrated as a validation step.

2 / 3

Progressive Disclosure

Excellent structure: concise overview and quick start in the main file, with clearly signaled one-level-deep references to hub-discovery.md, quantization.md, and hardware.md for detailed topics. The Resources section provides external links without cluttering the main content. Navigation is clear and well-organized.

3 / 3

Total

11

/

12

Passed

Description

100%

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

This is a strong skill description that clearly defines a specific niche (local model inference with llama.cpp/GGUF), lists concrete actions, and includes rich trigger terms covering hardware platforms and workflows. The 'Use to...' opening provides clear trigger guidance, and the domain-specific terminology makes it highly distinctive.

DimensionReasoningScore

Specificity

Lists multiple specific concrete actions: finding GGUFs, quant selection, running servers, exact GGUF file lookup, conversion, and OpenAI-compatible local serving. Also specifies hardware targets (CPU, Mac Metal, CUDA, ROCm).

3 / 3

Completeness

Clearly answers both what ('finding GGUFs, quant selection, running servers, exact GGUF file lookup, conversion, and OpenAI-compatible local serving') and when ('Use to select models to run locally with llama.cpp and GGUF on CPU, Mac Metal, CUDA, or ROCm'). The 'Use to...' clause serves as an explicit trigger.

3 / 3

Trigger Term Quality

Includes strong natural keywords users would say: 'llama.cpp', 'GGUF', 'local', 'CPU', 'Mac Metal', 'CUDA', 'ROCm', 'quant', 'OpenAI-compatible', 'local serving'. These cover the terms someone running local models would naturally use.

3 / 3

Distinctiveness Conflict Risk

Highly distinctive niche targeting llama.cpp, GGUF format, and specific hardware backends. Very unlikely to conflict with other skills given the specificity of the domain and terminology.

3 / 3

Total

12

/

12

Passed

Validation

100%

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation11 / 11 Passed

Validation for skill structure

No warnings or errors.

Repository
huggingface/context-course
Reviewed

Table of Contents

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.