cv-classification

Best practices for image classification tasks. Use when working on CIFAR, ImageNet, or other classification benchmarks.

Quality

76%

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 ./researchclaw/skills/builtin/domain/cv-classification/SKILL.md

Quality

Content

87%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 content is a tight, highly actionable reference that assumes competence and wastes no tokens. Its only gap is the absence of explicit validation/checkpoint guidance to turn the implied select-train-evaluate flow into a verified workflow.

Suggestions

Add a short validation checkpoint, e.g. 'After training, confirm val accuracy is within ~1-2% of the listed benchmark; if not, check lr schedule and augmentation.'

Consider an explicit ordered sequence (1. pick architecture by scale → 2. apply matching training recipe → 3. evaluate against benchmark) to make the implicit flow unambiguous.

Dimension	Reasoning	Score
Conciseness	The body is a lean set of bullet points with specific values and no padding or explanation of concepts Claude already knows; every line earns its place.	5 / 5
Actionability	Concrete, directly applicable guidance is given throughout — named architectures, exact lr/weight-decay ranges, specific augmentations, batch sizes, and benchmark accuracy targets — which is fully actionable for an instruction-only skill.	5 / 5
Workflow Clarity	Sections imply a loose sequence (architecture → training recipe → benchmark), but there are no explicit validation checkpoints or feedback loops to confirm a configuration is working as expected.	3 / 5
Progressive Disclosure	The body is under 50 lines, needs no external references, and is organized into three clearly labeled sections, satisfying the simple-skill progressive-disclosure criterion.	5 / 5
	Total	18 / 20 Passed

Description

65%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 pairs an explicit 'Use when…' trigger with recognizable CV benchmark keywords, giving it solid completeness and distinctiveness. Its main weakness is the generic 'Best practices for' framing, which under-specifies the concrete capabilities the skill provides.

Suggestions

Replace the generic 'Best practices for' framing with concrete actions, e.g. 'Select architectures, configure training recipes, and evaluate on CIFAR/ImageNet classification benchmarks.'

Surface a few of the metadata trigger keywords (CNN, ViT, vision) directly in the description to improve natural-term coverage.

Dimension	Reasoning	Score
Specificity	The phrase 'Best practices for image classification tasks' names the domain but offers only a minimal, generic action ('best practices') rather than concrete capabilities like architecture selection or training-recipe guidance.	2 / 5
Completeness	It answers both 'what' (best practices for image classification) and 'when' via an explicit 'Use when working on CIFAR, ImageNet, or other classification benchmarks' clause, but the 'what' is slightly soft rather than fully concrete.	4 / 5
Trigger Term Quality	'CIFAR, ImageNet, or other classification benchmarks' gives good natural keyword coverage that users would actually say, though common synonyms (CNN, ViT, vision) appear only in metadata keywords rather than the description.	4 / 5
Distinctiveness Conflict Risk	'Image classification tasks' with CIFAR/ImageNet triggers carves a mostly distinct niche with only minor overlap risk against adjacent CV skills like object detection or segmentation.	4 / 5
	Total	14 / 20 Passed

Validation

100%

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 — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository: aiming-lab/AutoResearchClaw
Path: researchclaw/skills/builtin/domain/cv-classification/SKILL.md
Commit: be4ba47

Reviewed: 5 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.