computer-vision-expert

SOTA Computer Vision Expert (2026). Specialized in YOLO26, Segment Anything 3 (SAM 3), Vision Language Models, and real-time spatial analysis.

Quality

26%

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 ./skills/computer-vision-expert/SKILL.md

Quality

Content

17%Scale 1-5

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

This skill reads as a high-level capability overview or role description rather than actionable instructions for Claude. It lacks any executable code, concrete commands, or specific implementation steps. The content is verbose with explanations of concepts Claude already knows (NMS, SLAM, VQA) while failing to provide the novel, specific guidance that would make it useful.

Suggestions

Add concrete, executable code examples for each major capability (e.g., a complete YOLO26 inference script, a SAM 3 text-to-mask pipeline, a depth estimation snippet with Depth Anything V2).

Replace the 'Capabilities' section's descriptive bullet points with actionable patterns showing specific API calls, model loading, and inference commands.

Add explicit multi-step workflows with validation checkpoints, especially for deployment pipelines (e.g., train → export ONNX → validate → convert TensorRT → benchmark).

Split detailed content into bundle files (e.g., YOLO26_DEPLOYMENT.md, SAM3_GUIDE.md, VLM_PATTERNS.md) and reference them from the main skill with clear navigation.

Dimension	Reasoning	Score
Conciseness	The skill extensively explains concepts and capabilities Claude would already know (what VQA is, what SLAM is, what NMS is). It reads more like a marketing brochure or capability overview than actionable instructions. Sections like 'Capabilities' and 'When to Use' are padded with descriptions rather than providing novel, token-efficient guidance.	2 / 5
Actionability	There is no executable code, no concrete commands, no specific API calls, and no copy-paste ready examples anywhere in the skill. Everything is abstract description ('Use SAM 3's text-to-mask capability', 'Leverage YOLO26's simplified ONNX/TensorRT exports') without showing how to actually do any of it.	1 / 5
Workflow Clarity	The 'Patterns' section hints at multi-step workflows (e.g., YOLO26 for candidate proposal then SAM 3 for mask refinement) but provides no concrete sequencing, no validation steps, and no error recovery. The steps are vague descriptions rather than actionable sequences.	2 / 5
Progressive Disclosure	The content is a monolithic file with no bundle files and no references to external detailed guides. All the capability descriptions are inlined without any structured navigation to deeper resources. The 'Related Skills' section lists skill names but provides no links or paths. For a skill this broad, content should be split across multiple files.	2 / 5
	Total	7 / 20 Passed

Description

36%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.

This description reads like a credential statement rather than a functional skill description. It lists technologies and buzzwords ('SOTA', '2026') but fails to describe what concrete actions the skill performs or when it should be triggered. The absence of a 'Use when...' clause and the lack of actionable verbs significantly weaken its utility for skill selection.

Suggestions

Add a 'Use when...' clause with natural trigger phrases like 'Use when the user asks to detect objects in images, segment image regions, analyze visual content, or work with YOLO/SAM models'.

Replace the credential-style opening ('SOTA Computer Vision Expert (2026)') with concrete action verbs describing what the skill does, e.g., 'Detects and localizes objects in images, performs instance segmentation, analyzes spatial relationships between objects'.

Include common user-facing synonyms and natural language terms such as 'object detection', 'image segmentation', 'bounding boxes', 'identify objects in photos', 'label images'.

Dimension	Reasoning	Score
Specificity	Names the domain (computer vision) and lists specific technologies (YOLO26, SAM 3, Vision Language Models), but does not describe concrete actions the skill performs. 'Real-time spatial analysis' is somewhat specific but still vague about what it actually does.	2 / 5
Completeness	Has a vague 'what' (specialized in certain technologies) but no 'when' clause at all. There is no explicit guidance on when Claude should select this skill. The description reads more like a resume headline than a functional skill description.	2 / 5
Trigger Term Quality	Includes some relevant keywords like 'YOLO26', 'SAM 3', 'Vision Language Models', 'computer vision', and 'spatial analysis', but these are mostly technical jargon rather than natural phrases users would say. Missing common user terms like 'object detection', 'image segmentation', 'detect objects in images', 'identify objects', etc.	3 / 5
Distinctiveness Conflict Risk	The specific technology names (YOLO26, SAM 3) provide some distinctiveness, but 'computer vision' and 'Vision Language Models' are broad enough to overlap with other vision-related skills. The lack of concrete actions makes it harder to distinguish from other CV-related skills.	3 / 5
	Total	10 / 20 Passed

Validation

90%

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 — 10 / 11 Passed

Validation for skill structure

Criteria	Description	Result
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	10 / 11 Passed

Repository: administrakt0r/AI-Agents-Safe-Coding-Skills
Path: skills/computer-vision-expert/SKILL.md
Commit: 57c135b

Reviewed: 2 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.