cv-detection

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

Quality

70%

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

72%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 is a compact, well-organized, actionable best-practices reference that respects token budget. Its main weakness is the absence of a sequenced workflow with validation checkpoints, and minor time-sensitive version references.

Suggestions

Move or qualify version/identifier references (YOLO v5/v8, COCO val2017) into a clearly labeled section to avoid time-sensitivity penalization, or omit specific versions.

If a training/evaluation workflow is intended, present it as an ordered numbered sequence with explicit validation checkpoints (e.g., verify mAP after each change).

Add at least one concrete, copy-paste-ready code snippet (e.g., mAP/IoU computation or a focal-loss config) to lift actionability from 4 to 5.

Dimension	Reasoning	Score
Conciseness	The body is lean bullet-point reference with no concept padding and assumes Claude's competence, near the anchor 5 example. It is capped at 4 because version/identifier references like "YOLO (v5/v8)" and "COCO val2017" are time-sensitive and not placed in a deprecated section.	4 / 5
Actionability	Concrete specifics are given (IoU 0.5 and 0.5:0.95, Focal loss, FPN, ImageNet backbone) but there is no executable code and some tips stay high-level ("Multi-scale training and testing"). It is not 5 because nothing is copy-paste ready, and not 3 because the guidance is specific and actionable rather than pseudocode.	4 / 5
Workflow Clarity	Content is organized into Architecture families / Training recipe / Standard benchmarks, but these are advisory tips rather than a sequenced procedure, with no validation checkpoints. It is not a destructive/batch skill so the cap-3 rule is not what forces this; it scores 3 because no ordered workflow or checkpoints are present.	3 / 5
Progressive Disclosure	The skill is under 50 lines, needs no external references, and is organized into three clearly headed sections with no nested references. Per the simple-skill exception, well-organized sections alone earn a 5, and no bundle files exist to require deeper structuring.	5 / 5
	Total	16 / 20 Passed

Description

68%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 clearly scopes a distinct niche and supplies explicit trigger terms, but its capability statement ("best practices") is generic and lacks concrete actions, capping specificity and completeness.

Suggestions

Replace the generic "Best practices for object detection tasks" with concrete actions, e.g. "Configure, train, and evaluate object detection models."

Add a few natural synonyms to the trigger clause such as "bounding boxes", "mAP", or "anchors" to broaden keyword coverage.

Keep the strong "Use when working on COCO, VOC, YOLO, or DETR" clause as-is; it is the strongest part of the description.

Dimension	Reasoning	Score
Specificity	The phrase "Best practices for object detection tasks" names the domain but the only action ("best practices") is generic, with no concrete verbs like extract/train/evaluate. It is not a 3 because there are no 1-2 concrete actions enumerated, only a vague label.	2 / 5
Completeness	Both a "what" (best practices for object detection) and an explicit "when" (Use when working on COCO/VOC/YOLO/DETR) are present. It is not a 5 because the "what" is generic rather than a concrete action list.	4 / 5
Trigger Term Quality	"Use when working on COCO, VOC, or detection architectures like YOLO and DETR" surfaces natural detection terms users would say. It is not a 5 because common synonyms such as "bounding box", "mAP", or "anchor" are absent from the description.	4 / 5
Distinctiveness Conflict Risk	Object detection is a clear niche and COCO/VOC/YOLO/DETR are distinct, unlikely-to-collide triggers, giving minimal conflict risk per the anchor.	5 / 5
	Total	15 / 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-detection/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.