axolotl

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

Quality

51%

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/training/axolotl/SKILL.md

Quality

Content

43%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 body provides useful concrete snippets and a real reference bundle, but is padded with boilerplate, contains non-executable bare-token code blocks, lacks any sequenced workflow, and references missing files. Tightening prose and aligning navigation with the actual bundle would help.

Suggestions

Remove or collapse the boilerplate "When to Use This Skill", "Working with This Skill", "Resources", and "Notes" sections that restate what Claude already infers.

Replace non-executable bare-token code blocks ("context_parallel_size", "integrations") with real, copy-pasteable YAML/commands or fold them into prose.

Fix navigation to match the actual bundle: remove references to non-existent getting_started/tutorials/guides files, and link the real reference files (api.md, dataset-formats.md, other.md) with one-line descriptions of when to consult each.

Dimension	Reasoning	Score
Conciseness	Core patterns are lean code snippets, but several boilerplate sections ("When to Use This Skill", "Working with This Skill", "Resources", "Notes") and a redundant context_parallel_size pair (Pattern 3/4) add padding Claude does not need.	3 / 5
Actionability	Mixes genuinely executable YAML/commands (FSDP config, save_compressed, all_reduce_perf) with non-executable blocks (bare tokens "context_parallel_size" and "integrations") and bare class/method signatures lacking usage context.	3 / 5
Workflow Clarity	This is an unsequenced grab-bag of standalone patterns with no real workflow; the only sequence is the trivial 2-step "Updating" process, which has gaps and no validation checkpoints.	2 / 5
Progressive Disclosure	A real one-level-deep bundle exists (api.md, dataset-formats.md, other.md, index.md) and is listed with descriptions, but the body points users to non-existent files ("getting_started", "tutorials", "guides") that are not in references/.	3 / 5
	Total	11 / 20 Passed

Description

58%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 and tool-anchored but terse and lacks any explicit "Use when..." trigger guidance, capping completeness. Adding a when-to-use clause and a few more method synonyms would lift it.

Suggestions

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

Expand the capability list from labels to concrete verb actions (e.g. "configure LoRA/QLoRA, run DPO/GRPO preference training, enable FSDP/DeepSpeed").

Include natural synonyms users say (QLoRA, preference optimization, RLHF) to improve trigger-term coverage.

Dimension	Reasoning	Score
Specificity	Names the domain and several concrete capabilities ("YAML LLM fine-tuning", "LoRA, DPO, GRPO"), but uses terse label-style method abbreviations rather than comprehensive verb-action coverage, leaving minor gaps (e.g. QLoRA/KTO/ORPO/multimodal appear in the body but not the description).	4 / 5
Completeness	Has a clear "what" (YAML LLM fine-tuning with LoRA/DPO/GRPO) but no "Use when..." or equivalent explicit trigger guidance, which per the rubric caps completeness at 3.	3 / 5
Trigger Term Quality	Includes relevant natural terms a user would say ("Axolotl", "LoRA", "DPO", "GRPO", "fine-tuning") but omits common variations and synonyms (QLoRA, KTO, ORPO, preference optimization) and any file-extension cues.	3 / 5
Distinctiveness Conflict Risk	The specific tool name "Axolotl" plus the YAML fine-tuning niche make it mostly distinct with only minor overlap risk against a generic LLM-training skill, though the terse phrasing keeps it short of fully disambiguating.	4 / 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/axolotl/SKILL.md
Commit: 6680afb

Reviewed: about 4 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.