Back to rankings

NVIDIA / SkillEvaluator

Multi-tier framework for evaluating AI agent skills with quality gates, semantic overlap detection, synthetic evaluation dataset generation, and live agent evaluation that measures how skills affect agent behavior.

Repository stars
646All-time #52,719
Fork
84
Last push
updated 1 days ago
Repository created
since 2026
Apps & systemsGeneralPythonApache-2.0 · Permissive
View sourceWebsite

About

Original · GitHub
Multi-tier framework for evaluating AI agent skills with quality gates, semantic overlap detection, synthetic evaluation dataset generation, and live agent evaluation that measures how skills affect agent behavior.

Summarised by AI from the README; the original prevails

Author description: Multi-tier framework for evaluating AI agent skills with quality gates, semantic overlap detection, synthetic evaluation dataset generation, and live agent evaluation that measures how skills affect agent behavior.

Basic listing based on official public information; functionality has not been independently evaluated.

Pricing
Pricing not verifiedCheck current official software, hosting, model and infrastructure pricing.

Star trend

64664764711-06 · 64610-0810-1310-1810-2310-2811-0211-06

First collected 2026-11-06Measured at 2026-11-06 07:51 UTCSourceLicense source

Compiled from public project information; not independently tested. · Repository Stars, not unique users · License tiers are derived from metadata. Check obligations for your use; unknown does not mean prohibited.