Name: roboflow-training-and-evaluation
Rating: 69.6 (1 reviews)
Author: roboflow

roboflow-training-and-evaluation

Use when training Roboflow models, improving accuracy, or setting up a production feedback loop — covers architecture selection, model IDs, checkpoints, evaluation metrics, the iterative improvement playbook, and Active Learning through the Project Model Workflow block.

Quality

83%

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

Quality

Content

75%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 well-structured, highly actionable reference skill with exact model IDs, a branching decision tree, and explicit error-handling guidance. It is consistently strong but settles at 4 across dimensions due to minor verbosity, absence of full executable code blocks, and inlined detail that could be split out.

Suggestions

Add one or two complete copy-paste example blocks (e.g. a start-training + poll-status + read-NAS-children sequence) to push actionability toward fully executable.

Move the large inlined reference material (COCO 80 class list, full RF-DETR NAS deep-dive) into separate referenced files and link to them one level deep to improve progressive disclosure.

Tighten the opening source-of-truth blockquote and the NAS introductory prose to remove over-explanation and lift conciseness.

Dimension	Reasoning	Score
Conciseness	Dense reference tables and exact model IDs are token-efficient, but the source-of-truth blockquote and a few explanatory prose passages (e.g. the NAS intro) could be trimmed, keeping it just below lean.	4 / 5
Actionability	Provides exact model_id values, tool-call signatures with parameters, and specific error codes to branch on, but offers no full copy-paste executable code blocks, so it stops short of the top anchor.	4 / 5
Workflow Clarity	A 14-step decision tree with explicit branching plus NAS prerequisite checks and error-code fallbacks gives a clear sequenced workflow; post-training validation is mostly auto/implicit rather than an explicit checklist, a minor gap.	4 / 5
Progressive Disclosure	Clear section headers and two well-signaled one-level-deep Related Pages links, but a large amount of reference material (architecture tables, COCO 80 list, full NAS deep-dive) is inlined rather than split into referenced files.	4 / 5
	Total	16 / 20 Passed

Description

92%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 with an explicit Use-when trigger and a clear enumeration of capabilities scoped to the Roboflow training/evaluation niche. Its only gap is keyword synonym coverage, which keeps trigger_term_quality just below the top anchor.

Dimension	Reasoning	Score
Specificity	Enumerates six concrete capabilities (architecture selection, model IDs, checkpoints, evaluation metrics, improvement playbook, Active Learning), matching the comprehensive-coverage anchor.	5 / 5
Completeness	Explicitly answers both 'what' (covers architecture selection, model IDs, checkpoints, metrics, playbook, Active Learning) and 'when' (a clear 'Use when...' clause with concrete triggers).	5 / 5
Trigger Term Quality	Has natural phrases ('training Roboflow models', 'improving accuracy', 'production feedback loop') but misses common synonyms users would say like 'fine-tune', 'retrain', or 'evaluation', so it falls short of comprehensive.	4 / 5
Distinctiveness Conflict Risk	Tightly scoped to the Roboflow training/evaluation niche with product-specific triggers ('Project Model Workflow block', 'Active Learning'), giving minimal conflict risk with other skills.	5 / 5
	Total	19 / 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: roboflow/computer-vision-skills
Path: skills/training-and-evaluation/SKILL.md
Commit: f3b2cbd

Reviewed: 10 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.