llava

Vision-language chat: VQA, captioning, image dialogue.

Quality

53%

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/llava/SKILL.md

Quality

Content

53%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 solid executable quick-start and CLI examples but is over-long, inlines content that belongs in the existing reference file, and lacks validation checkpoints for batch/destructive operations.

Suggestions

Replace the inlined 'Training custom model' section with a one-level reference to references/training.md, e.g. 'See [training.md](references/training.md) for training stages and configs.'

Add validation/checkpoint steps to batch image processing (e.g. verify image loaded, confirm output non-empty before proceeding) to lift workflow clarity above 3.

Remove marketing metrics, benchmark tables, and the trivial 'Common tasks' one-liners to tighten the token budget.

Dimension	Reasoning	Score
Conciseness	The body is mostly efficient with real code, but is padded with marketing metrics ('23,000+ GitHub stars', benchmark tables) and trivial one-line 'Common tasks' sections that could be trimmed, fitting 'mostly efficient but includes some unnecessary explanation'.	3 / 5
Actionability	Quick start, CLI, and Gradio examples are concrete and executable, but the 'Common tasks' section relies on an undefined `ask()` helper and the LangChain example is a stub returning `response`, leaving minor gaps.	4 / 5
Workflow Clarity	Multi-turn conversation and two training stages are sequenced, but there are no validation/verification checkpoints and the rubric caps batch/destructive operations without validation at 3.	3 / 5
Progressive Disclosure	A `references/training.md` bundle file exists but is never linked from the body, and training content is inlined instead of referenced, fitting 'content that clearly belongs in separate files is inlined'.	2 / 5
	Total	12 / 20 Passed

Description

53%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 clear niche with concrete tasks, but it lacks any explicit 'when to use' trigger guidance and relies on abbreviations rather than the natural phrases users would say.

Suggestions

Add an explicit 'Use when...' clause, e.g. 'Use when the user wants to chat about, describe, or ask questions about an image.'

Expand abbreviated tasks into natural phrasings users actually say ('visual question answering', 'describe this image', 'image captioning').

Include common synonyms or file/format cues to widen trigger coverage and reduce overlap with generic image skills.

Dimension	Reasoning	Score
Specificity	Names the domain ('vision-language chat') and three concrete tasks ('VQA, captioning, image dialogue'), matching the 'names domain and 1-2 concrete actions' anchor; it is not a 4 because the actions are abbreviated and not comprehensive.	3 / 5
Completeness	It gives a clear 'what' but no 'when'/'Use when' trigger guidance, so per the rubric a missing 'Use when' clause caps completeness at 3.	3 / 5
Trigger Term Quality	'VQA' and 'captioning' are relevant natural terms, but common phrasings users would actually say (e.g. 'describe this image', 'ask about a picture') are missing, so it lands at 'some relevant keywords but missing common variations'.	3 / 5
Distinctiveness Conflict Risk	'Vision-language chat' is a distinct niche with clear triggers and only minor overlap risk with general image skills, fitting 'mostly distinct; minor overlap risk'.	4 / 5
	Total	13 / 20 Passed

Validation

75%

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 — 12 / 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
referenced_paths_exist	Referenced path issues: 2 missing, 2 deeper-than-1-level	Warning

	Total	12 / 16 Passed

Repository: NousResearch/hermes-agent
Path: optional-skills/mlops/llava/SKILL.md
Commit: 6680afb

Reviewed: about 3 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.