roboflow-universe
Use when searching for or using public datasets/models on Roboflow Universe (universe.roboflow.com), the open repository of 1M+ computer vision datasets and 50K+ pre-trained models.
78
70%
Does it follow best practices?
Impact
89%
1.23xAverage score across 3 eval scenarios
Advisory
Suggest reviewing before use
Optimize this skill with Tessl
npx tessl skill review --optimize ./skills/universe/SKILL.mdQuality
Discovery
89%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 is a solid description with clear trigger guidance and strong distinctiveness through platform-specific references. Its main weakness is that the 'what' portion could be more specific about the concrete actions the skill enables beyond 'searching for or using'. The trigger terms are well-chosen and cover natural user language around computer vision datasets and models.
Suggestions
Add more specific concrete actions such as 'browse, download, and deploy datasets/models' or 'run inference with pre-trained models, export annotations, explore dataset statistics' to improve specificity.
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | The description names the domain (public datasets/models on Roboflow Universe) and mentions searching and using, but doesn't list specific concrete actions like downloading datasets, running inference, browsing model architectures, etc. | 2 / 3 |
Completeness | Explicitly answers both 'what' (searching for/using public datasets and models on Roboflow Universe) and 'when' (starts with 'Use when' clause specifying the trigger conditions). The 'when' is clearly stated upfront. | 3 / 3 |
Trigger Term Quality | Includes strong natural trigger terms: 'datasets', 'models', 'Roboflow Universe', 'universe.roboflow.com', 'computer vision', 'pre-trained models', 'open repository'. These cover terms users would naturally use when looking for CV datasets or models. | 3 / 3 |
Distinctiveness Conflict Risk | Highly distinctive with specific platform references (Roboflow Universe, universe.roboflow.com) and a clear niche (computer vision datasets/models repository). Unlikely to conflict with other skills. | 3 / 3 |
Total | 11 / 12 Passed |
Implementation
50%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This is a solid reference skill that covers Roboflow Universe comprehensively with useful tables for query operators, licenses, and evaluation criteria. Its main weaknesses are the lack of executable code examples (relying on UI descriptions and MCP tool descriptions without example calls/responses), missing validation checkpoints in workflows, and a somewhat monolithic structure that could benefit from splitting detailed reference content into separate files.
Suggestions
Add concrete executable examples: show a sample `universe_search` MCP tool call with example input parameters and a truncated example response JSON to make the tool usage copy-paste ready.
Add validation/verification steps to the forking workflow (e.g., 'Verify the fork appears in your workspace project list' and 'Check that class counts match the source dataset').
Remove or significantly shorten the plugin/MCP preamble blockquote at the top — it's meta-information about skill loading that doesn't help Claude perform the task and consumes tokens.
Consider splitting the query operators reference table and license table into a separate REFERENCE.md file, keeping SKILL.md as a concise overview with links to detailed references.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The skill is generally well-structured and avoids explaining basic concepts, but includes some unnecessary content like the verbose plugin/MCP preamble at the top (which is meta-information about skill loading, not actionable guidance) and some tables that could be tighter. The license table and evaluation criteria are useful but slightly verbose. | 2 / 3 |
Actionability | The skill provides concrete query operators, URL patterns, and MCP tool parameters, which are highly actionable. However, it lacks executable code examples (e.g., actual API calls, Python snippets for downloading/searching programmatically). The forking and downloading sections describe UI steps rather than providing code. The MCP tool reference is concrete but no example call/response is shown. | 2 / 3 |
Workflow Clarity | Multi-step processes like forking and using Universe models are listed as sequences, but lack validation checkpoints. For example, the forking workflow doesn't mention verifying the fork completed successfully or checking class mappings. The 'Evaluating a Dataset' section is a good checklist but isn't integrated into a workflow with decision points or error recovery. | 2 / 3 |
Progressive Disclosure | The skill references related skills at the bottom (data-management, training-and-evaluation) which is good, but the document itself is fairly long and monolithic. The query operators, evaluation criteria, licenses, and model usage could potentially be split into separate reference files. No bundle files are provided to support progressive disclosure, and the content is all inline. | 2 / 3 |
Total | 8 / 12 Passed |
Validation
100%Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.
Validation — 11 / 11 Passed
Validation for skill structure
No warnings or errors.
- Repository
- roboflow/computer-vision-skills
- Commit
02936d5
- Reviewed
Table of Contents
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.