llava

Large Language and Vision Assistant. Enables visual instruction tuning and image-based conversations. Combines CLIP vision encoder with Vicuna/LLaMA language models. Supports multi-turn image chat, visual question answering, and instruction following. Use for vision-language chatbots or image understanding tasks. Best for conversational image analysis.

Quality

67%

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 ./skills/multimodal/llava/SKILL.md

Quality

Content

56%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.

A comprehensive, mostly executable LLaVA usage guide with clear sequencing, undermined by pseudocode stubs in several sections and a missing link to the existing training reference file.

Suggestions

Define or replace the undefined ask() and generate() helpers used in the multi-turn and Common tasks sections with the actual executable inference code from the quick-start example.

Link to references/training.md from the "Training custom model" section instead of only inlining the two shell commands, so the bundled reference is discoverable.

Collapse the five repetitive "Common tasks" stubs into a single short example showing how to vary the question string, removing the padded subsections.

Dimension	Reasoning	Score
Conciseness	The body avoids explaining concepts Claude already knows, but the five near-identical "Common tasks" stubs and the lengthy repetitive sections add padding that could be tightened.	3 / 5
Actionability	The quick-start, CLI, quantization, and training examples are executable, but several sections rely on undefined helpers (ask(), generate()) and the LangChain LLaVALLM stub returns an undefined response, leaving key gaps.	3 / 5
Workflow Clarity	The install-to-load-to-generate sequence and the staged training flow are clearly ordered with minor gaps, and there are no destructive or batch operations that would demand explicit validation checkpoints.	4 / 5
Progressive Disclosure	Section headers provide structure, but the bundled references/training.md is never linked from the body and substantial content (benchmarks, integrations, common tasks) is inlined that could live in separate files.	3 / 5
	Total	13 / 20 Passed

Description

78%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.

A strong, specific description that covers concrete capabilities, architecture, and explicit use-case triggers with low conflict risk. The main improvement is adding more natural synonyms and concrete user-mention trigger phrases.

Dimension	Reasoning	Score
Specificity	Lists several concrete capabilities ("visual instruction tuning", "image-based conversations", "multi-turn image chat", "visual question answering", "instruction following") rather than vague language, with only minor coverage gaps.	4 / 5
Completeness	Clearly states both what it does (vision-language model combining CLIP and Vicuna/LLaMA) and when to use it ("Use for vision-language chatbots or image understanding tasks"), but the trigger could name user-mention phrases more concretely.	4 / 5
Trigger Term Quality	Includes natural user-facing phrases like "image chat", "visual question answering", and "image understanding", though it misses common synonyms and file extensions that would push it to a 5.	4 / 5
Distinctiveness Conflict Risk	Names the LLaVA-specific architecture ("Combines CLIP vision encoder with Vicuna/LLaMA language models") giving it a clear niche with minimal overlap risk against generic image skills.	5 / 5
	Total	17 / 20 Passed

Validation

87%

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 — 14 / 16 Passed

Validation for skill structure

Criteria	Description	Result
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	14 / 16 Passed

Repository: OpenLAIR/dr-claw
Path: skills/multimodal/llava/SKILL.md
Commit: 4b35756

Reviewed: 1 day 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.