axolotl

Axolotl: YAML LLM fine-tuning (LoRA, DPO, GRPO).

Quality

57%

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/axolotl/SKILL.md

Quality

Content

57%

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

The content is well-structured for progressive disclosure, pointing to real reference files for detail while keeping an overview inline. It is weakened by generic boilerplate sections and several incomplete 'patterns' that present bare tokens instead of executable examples.

Suggestions

Remove generic boilerplate (the 'Notes', 'Updating', and auto-generation disclaimer lines) and tighten Patterns 3-4 into one concrete, runnable example.

Replace bare-token code blocks like `context_parallel_size` and `integrations` with actual working snippets or explicit prose so each Quick Reference entry is copy-paste ready.

Either add a brief sequenced workflow with a validation/check step for a representative fine-tuning task, or reframe the guidance so the single discovery action (consult the right reference file) is stated unambiguously.

Dimension	Reasoning	Score
Conciseness	Concrete patterns and code earn their place, but boilerplate like "Comprehensive assistance with axolotl development, generated from official documentation" plus generic Notes/Updating/Working-with sections and redundant Patterns 3-4 add padding, matching the 'mostly efficient but could be tightened' anchor.	2 / 3
Actionability	Several examples are executable (the FSDP YAML block, the NCCL test command, the Python API signatures), but Patterns 3, 4, and 6 reduce to bare tokens like `context_parallel_size` and `integrations` as code blocks without a complete runnable example, landing at 'some concrete guidance but incomplete'.	2 / 3
Workflow Clarity	The 'Working with This Skill' section offers loose pointers (start with getting_started, use category files) but no sequenced multi-step process or validation checkpoints, so it sits at 'steps listed but checkpoints missing/implicit' rather than a fully sequenced workflow.	2 / 3
Progressive Disclosure	The body is an overview that clearly signals one-level-deep reference files (api.md, dataset-formats.md, other.md) which all exist under references/, with detail appropriately split out, matching the 'clear overview with well-signaled one-level-deep references' anchor.	3 / 3
	Total	9 / 12 Passed

Description

57%

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 distinctive and reasonably specific, naming the tool and several fine-tuning methods. It is held back by the absence of an explicit 'Use when...' trigger clause and by incomplete coverage of the method variants the skill actually supports.

Suggestions

Add an explicit 'Use when...' clause, e.g. 'Use when fine-tuning LLMs with Axolotl YAML configs or when the user mentions LoRA/QLoRA, DPO/KTO/ORPO, or GRPO training.'

Include the missing method variants (QLoRA, KTO, ORPO) so trigger-term coverage matches the skill's actual capabilities.

Lead with a concrete verb phrase rather than a label (e.g. 'Fine-tune LLMs via Axolotl YAML configs...') to strengthen specificity.

Dimension	Reasoning	Score
Specificity	Quotes "YAML LLM fine-tuning (LoRA, DPO, GRPO)" name the domain and list concrete methods, but the action set is narrow (only fine-tuning) and not comprehensive, matching the 'names domain and some actions' anchor rather than the multi-action level-3 example.	2 / 3
Completeness	It states what the skill does (YAML LLM fine-tuning) but lacks any explicit 'Use when...' trigger clause, which per the guidelines caps completeness at 2 rather than 3.	2 / 3
Trigger Term Quality	Terms like "Axolotl", "LLM fine-tuning", "LoRA", "DPO", "GRPO" are natural, but common variations the skill itself covers (QLoRA, KTO, ORPO) are omitted, so it falls to 'some relevant keywords but missing common variations' rather than full coverage.	2 / 3
Distinctiveness Conflict Risk	The named tool "Axolotl" plus a narrow fine-tuning niche make it clearly distinguishable from other skills and unlikely to trigger for the wrong skill, matching the clear-niche level-3 anchor.	3 / 3
	Total	9 / 12 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/axolotl/SKILL.md
Commit: 4c9628e

Reviewed: about 15 hours 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.