unsloth

Unsloth: 2-5x faster LoRA/QLoRA fine-tuning, less VRAM.

Quality

39%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Fix and improve this skill with Tessl

tessl review fix ./optional-skills/mlops/training/unsloth/SKILL.md

Quality

Content

21%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 skill body is a thin auto-generated shell dominated by placeholders and generic boilerplate, with no executable guidance and no real workflow. It points to reference files but only signals one of the four present, leaving discovery poor.

Suggestions

Replace the empty "Quick Reference" placeholder with 2-3 concrete, copy-pasteable unsloth code patterns (e.g. FastModel load + LoRA training loop).

Document a real multi-step fine-tuning workflow with explicit validation checkpoints (load model, attach LoRA, train, save adapter, verify loss/generation).

List and label all reference files actually present (index.md, llms-txt.md, llms-full.md, llms.md) with one-line descriptions so Claude can navigate them, and trim the generic Notes/Updating boilerplate.

Dimension	Reasoning	Score
Conciseness	The body is short but largely generic boilerplate ("Reference files preserve the structure...", "Code examples include language detection...") and placeholder sections ("patterns will be added as you use the skill") that add little value Claude does not already infer.	2 / 5
Actionability	There is no executable code, no concrete commands, and no specific examples; the only actionable instruction is "Use `view` to read specific reference files", which is abstract direction rather than instruction.	1 / 5
Workflow Clarity	Only a rough, vague pointer structure exists ("Start with getting_started", "use the appropriate category reference file") with no real sequence and no validation checkpoints for the fine-tuning workflow.	2 / 5
Progressive Disclosure	The body is an overview pointing to reference files (good in principle), but it only signals one of the four actual bundle files (llms-txt.md) while index.md, llms-full.md, and llms.md exist unmentioned, so navigation is incomplete.	3 / 5
	Total	8 / 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 concise and names a distinct, well-scoped niche, but it lacks an explicit trigger clause and only describes capability properties rather than concrete actions. Adding a "Use when..." phrase would lift completeness and trigger-term quality.

Suggestions

Append a trigger clause, e.g. "Use when fine-tuning or training LLMs with LoRA/QLoRA, reducing VRAM usage, or speeding up adapter-based training."

Add natural synonyms users say ("PEFT", "adapters", "language model training") to broaden trigger-term coverage.

Mention a concrete action or two (e.g. "configure LoRA adapters", "launch faster training runs") instead of only stating performance properties.

Dimension	Reasoning	Score
Specificity	Names the domain ("LoRA/QLoRA fine-tuning") with two concrete capability claims ("2-5x faster", "less VRAM"), but does not enumerate multiple discrete actions a user would perform.	3 / 5
Completeness	Clearly states what the skill does (faster LoRA/QLoRA fine-tuning, less VRAM) but provides no "Use when..." clause or equivalent explicit trigger guidance, which caps completeness at 3 per the rubric.	3 / 5
Trigger Term Quality	Includes relevant domain keywords (LoRA, QLoRA, fine-tuning, VRAM, Unsloth) but omits common natural variations users say such as "training", "PEFT", "adapters", or "language model".	3 / 5
Distinctiveness Conflict Risk	Names a specific tool (Unsloth) coupled with specific techniques (LoRA/QLoRA), giving it a clear niche with minimal overlap risk against other skills.	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/training/unsloth/SKILL.md
Commit: 6327930

Reviewed: 3 days ago

Table of Contents

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.