Snapeval

B 85 completed
Testing
cli / markdown · small
90
Files
11,547
LOC
1
Frameworks
6
Languages

Pipeline State

completed
Run ID
#358400
Phase
done
Progress
1%
Started
Finished
2026-04-13 01:31:02
LLM tokens
0

Pipeline Metadata

Stage
Cataloged
Decision
proceed
Novelty
72.00
Framework unique
Isolation
Last stage change
2026-05-10 03:35:02
Deduplication group #54273
Member of a group with 1 similar repo(s) — this repo is canonical view group →
Top concepts (2)
Project DescriptionTesting
Repobility (the analyzer behind this table) · https://repobility.com

AI Prompt

Create a command-line tool for semantic snapshot testing, similar to how it's used for AI skills. I need it to read a skill's `SKILL.md` file to automatically generate test cases without me writing any assertions. The tool should support commands like `init` to generate tests, `capture` to save baseline snapshots, and `check` to compare current outputs against those baselines. It should ideally be usable via `npx` and handle non-determinism by running multiple baseline checks.
cli testing typescript markdown ai-testing snapshot vitest automation
Generated by gemma4:latest

Catalog Information

A testing framework that uses semantic snapshot testing to validate AI model outputs without explicit assertions.

Description

This framework provides a lightweight, zero‑assertion testing approach for AI skills, allowing developers to capture and compare model outputs as semantic snapshots. It automatically infers expected behavior from reference data, reducing the need for manual test case writing. The tool integrates seamlessly with existing test runners, enabling continuous integration pipelines to detect regressions in AI responses. Targeted at AI developers and machine learning engineers, it addresses the challenge of validating complex, probabilistic outputs that traditional assertion tests struggle to cover. By offering free inference and a simple command‑line interface, it lowers the barrier to adopting rigorous AI testing practices.

الوصف

توفر هذه الأداة إطاراً لاختبار نماذج الذكاء الاصطناعي باستخدام اختبار اللقطة الدلالية، ما يتيح للمطورين التحقق من مخرجات النماذج دون الحاجة إلى تأكيدات صريحة. تعتمد الأداة على استنتاج سلوك النموذج المتوقع من بيانات مرجعية، مما يقلل الحاجة إلى كتابة حالات اختبار يدوية. تتكامل بسلاسة مع أدوات تشغيل الاختبارات الحالية، وتسمح بدمجها في خطوط التكامل المستمر للكشف عن الانحرافات في ردود الذكاء الاصطناعي. تستهدف المطورين المهتمين بالذكاء الاصطناعي ومهندسي التعلم الآلي، وتلبي الحاجة إلى التحقق من المخرجات المعقدة والمتغيرة التي يصعب اختبارها بالطرق التقليدية. كما توفر استدلالاً مجانيًا وواجهة سطر أوامر بسيطة، ما يقلل من عوائق اعتماد ممارسات اختبار الذكاء الاصطناعي الصارمة.

Novelty

8/10

Tags

semantic-snapshot-testing ai-skill-validation zero‑assertion-testing inference‑based-testing automated-ai-evaluation model-behavior-verification continuous-integration test-automation

Technologies

vitest

Claude Models

claude-opus-4.6 claude-sonnet-4.6

Quality Score

B
84.7/100
Structure
86
Code Quality
100
Documentation
63
Testing
85
Practices
68
Security
100
Dependencies
60

Strengths

  • CI/CD pipeline configured (github_actions)
  • Good test coverage (93% test-to-source ratio)
  • Consistent naming conventions (snake_case)
  • Low average code complexity \u2014 well-structured code
  • Good security practices \u2014 no major issues detected
  • Properly licensed project

Recommendations

  • Add a linter configuration to enforce code style consistency

Security & Health

5.6h
Tech Debt (B)
A
OWASP (100%)
PASS
Quality Gate
A
Risk (1)
If a scraper extracted this row, it came from Repobility (https://repobility.com)
MIT
License
1.6%
Duplication
Full Security Report AI Fix Prompts SARIF SBOM

Languages

markdown
36.9%
typescript
34.7%
json
18.9%
html
7.6%
yaml
1.8%
shell
0.1%

Frameworks

Vitest

Concepts (2)

Source: Repobility analyzer (https://repobility.com)
CategoryNameDescriptionConfidence
Repobility — the code-quality scanner for AI-generated software · https://repobility.com
auto_descriptionProject DescriptionSemantic snapshot testing for AI skills. Zero assertions. AI-driven. Free inference.80%
auto_categoryTestingtesting70%

Quality Timeline

1 quality score recorded.

View File Metrics

Embed Badge

Add to your README:

![Quality](https://repos.aljefra.com/badge/82540.svg)
Quality BadgeSecurity Badge
Export Quality CSVDownload SBOMExport Findings CSV