cv-detection

Best practices for object detection tasks. Use when working on COCO, VOC, or detection architectures like YOLO and DETR.

Quality

73%

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-detection/SKILL.md

Quality

Content

57%Weight 40%Scale 1-3

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

A compact, well-organized best-practices reference that assumes Claude's competence and needs no external files. It is held back by time-sensitive version/benchmark figures, the absence of executable guidance, and a lack of any sequenced workflow with validation checkpoints.

Suggestions

Move or isolate time-sensitive figures (YOLO v5/v8, '~37 mAP', '~51 mAP') into a clearly labeled benchmarks/versions section so stale numbers don't dilute the evergreen training recipe.

Add at least one concrete, executable artifact — e.g. a sample training command or a short config snippet — so the 'Training recipe' is copy-paste ready rather than bullet-level tips.

If a real workflow is intended (e.g. train -> evaluate on COCO val2017 -> compare mAP), sequence it as numbered steps with a validation/checkpoint step; otherwise explicitly frame the body as a reference checklist.

Dimension	Reasoning	Score
Conciseness	The body is lean and avoids explaining basics Claude already knows, but it embeds time-sensitive version numbers ('YOLO (v5/v8)') and benchmark figures ('~37 mAP', '~51 mAP') that will date and are not placed in a deprecated/old-patterns section, which the rubric penalizes.	2 / 3
Actionability	Recommendations are specific and actionable ('IoU threshold: 0.5 for mAP50, 0.5:0.95 for mAP', 'Focal loss for class imbalance'), but they remain high-level tips with no executable commands, configs, or code, fitting the 'concrete but incomplete' level-2 anchor.	2 / 3
Workflow Clarity	Content is organized into clear categories (architecture families, training recipe, benchmarks), but there is no sequenced multi-step workflow or validation checkpoint — appropriate for a reference list, but it does not reach the explicit-checkpoint level-3 anchor.	2 / 3
Progressive Disclosure	At under 50 lines with no bundle files present and no external references needed, the well-organized sectioned overview meets the rubric's allowance for simple skills to score 3 on progressive disclosure without separate files.	3 / 3
	Total	9 / 12 Passed

Description

90%Weight 40%Scale 1-3

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 concise, third-person description with an explicit 'Use when' trigger and strong, natural domain keywords that give it a clear niche. Its only weakness is the abstract 'best practices' phrasing, which names no concrete actions.

Dimension	Reasoning	Score
Specificity	It names the domain ('object detection tasks') and concrete architectures/datasets (COCO, VOC, YOLO, DETR), but the only action verb is the vague 'best practices' — no concrete capabilities are enumerated, so it sits below the multi-action level-3 anchor yet above 'Helps with documents'.	2 / 3
Completeness	It states both what ('Best practices for object detection tasks') and an explicit when ('Use when working on COCO, VOC, or detection architectures like YOLO and DETR'), matching the level-3 anchor with an explicit 'Use when' trigger clause.	3 / 3
Trigger Term Quality	The trigger terms 'COCO, VOC, or detection architectures like YOLO and DETR' are exactly the natural keywords a practitioner would say when they need this skill, giving good coverage rather than just one term.	3 / 3
Distinctiveness Conflict Risk	Object detection scoped to COCO/VOC/YOLO/DETR is a clear niche with distinct triggers unlikely to fire for unrelated skills, matching the level-3 'clear niche with distinct triggers' anchor.	3 / 3
	Total	11 / 12 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-detection/SKILL.md
Commit: e2e23c9

Reviewed: 6 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.