Tessl • The package manager for agent skills

Blog Docs Log inGet started

SkillPublic

llama-cpp

llama.cpp local GGUF inference + HF Hub model discovery.

Quality

78%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Low

Low-risk findings worth noting

Fix and improve this skill with Tessl

tessl review fix ./optional-skills/mlops/inference/llama-cpp/SKILL.md

Quality

Content

92%Weight 40%Scale 1-5

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

The content is a strong, actionable skill body: executable commands and complete code throughout, a discovery workflow with explicit validation and fallback checkpoints, and clean progressive disclosure into six real reference files. The only weakness is minor redundancy (repeated tree API URL, worked example paragraph, long Resources list) that could be tightened.

Dimension	Reasoning	Score
Conciseness	The body is command- and URL-driven with almost no conceptual padding, but has minor trim opportunities: the tree API URL appears in three separate sections, and the worked Qwen3.6 example paragraph plus the eight-link Resources section repeat information available elsewhere. This matches 'efficient; minor instances of over-explanation that could be trimmed' rather than the every-token-earns-its-place anchor at 5.	4 / 5
Actionability	Every section gives copy-paste-ready material: 'brew install llama.cpp', 'llama-cli -hf bartowski/Llama-3.2-3B-Instruct-GGUF:Q8_0', a complete curl request for the OpenAI-compatible endpoint, full runnable Python snippets, and exact URL shapes for discovery. This matches 'fully executable; copy-paste ready code or commands'.	5 / 5
Workflow Clarity	The Model Discovery workflow is a clearly sequenced 7-step process with explicit validation and fallback loops: check the local-app snippet and 'if that section is not visible... say so and fall back to the tree API', 'query the tree API to confirm what actually exists', and only suggest conversion 'if the repo does not already expose GGUF files'. These error-recovery checkpoints match the top anchor.	5 / 5
Progressive Disclosure	The body is a well-organized overview (quick start, quant guidance, discovery workflow, output format) with six real, one-level-deep reference files clearly signaled in a References section, each with an accurate one-line description. Verified all six referenced files exist in references/ with content matching their labels.	5 / 5
	Total	19 / 20 Passed

Description

57%Weight 40%Scale 1-5

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

The description is specific to its niche and unlikely to conflict with other skills, but it reads as a capability summary rather than a trigger description. It lacks a 'Use when...' clause and misses natural trigger variations (Hugging Face spelled out, quantization terms).

Suggestions

Add an explicit trigger clause, e.g. 'Use when the user wants to run local GGUF models, pick a quant for their RAM/VRAM, or find GGUF files on Hugging Face.'

Spell out 'Hugging Face' instead of only 'HF Hub' and add quantization trigger terms ('quant', 'Q4_K_M', 'quantize') since quant selection is a core use case.

List one or two more concrete actions (e.g. 'build llama-server commands' or 'choose Q4/Q5/Q6/IQ quants') to round out capability coverage.

Dimension	Reasoning	Score
Specificity	The description names the domain ('llama.cpp', 'GGUF', 'HF Hub') and two concrete actions ('local GGUF inference', 'model discovery'), matching the '1-2 concrete actions, but not comprehensive' anchor. It is not a 4 because it omits several specific capabilities the skill actually covers (quant selection, server launch, Python bindings).	3 / 5
Completeness	The 'what' is clear (local GGUF inference, Hub model discovery), but there is no 'Use when...' clause or equivalent explicit trigger guidance, which caps completeness at 3 per the judging guidelines. It is not a 2 because the 'what' half is concrete, not vague.	3 / 5
Trigger Term Quality	'llama.cpp' and 'GGUF' are natural terms users would say, but the description abbreviates 'Hugging Face' to 'HF Hub' and omits common variations like 'quantize/quantization', 'run a local model', or 'llama-server'. This fits 'some relevant keywords but missing common variations or synonyms' rather than the good-coverage anchor at 4.	3 / 5
Distinctiveness Conflict Risk	'llama.cpp local GGUF inference + HF Hub model discovery' occupies a clear niche with distinct triggers (llama.cpp, GGUF); it is unlikely to fire for unrelated skills, matching the minimal-conflict anchor.	5 / 5
	Total	14 / 20 Passed

Validation

81%

Warnings & errors only

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

Validation — 13 / 16 Passed

Validation for skill structure

Criteria	Description	Result
metadata_version	'metadata.version' is missing	Warning
metadata_field	'metadata' should map string keys to string values	Warning
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	13 / 16 Passed

Repository: NousResearch/hermes-agent
Path: optional-skills/mlops/inference/llama-cpp/SKILL.md
Revision: ed728c5

Reviewed: 11 days ago

Discovery Implementation Validation

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.

llama-cpp

Table of Contents

Is this your skill?