Skip to content

Multi-tier framework for evaluating AI agent skills and plugins with quality gates, security scanning, harness compatibility checks, semantic overlap detection, synthetic evaluation dataset generation, and live agent evaluation that measures how skills affect agent behavior. Ships reusable CI templates that bring every check to any repository.

1.3 KB
Raw Blame
cff-version: 1.2.0
message: "If you use SkillEvaluator in academic or technical work, please cite the methodology paper under preferred-citation and record the SkillEvaluator version used for reproducibility."
title: "SkillEvaluator"
type: software
version: "1.0.1"
authors:
  - name: "NVIDIA Corporation"
license: Apache-2.0
repository-code: "https://github.com/NVIDIA/SkillEvaluator"
preferred-citation:
  type: article
  title: "Evaluating Skills, Not Just Agents: Agentic Continuous Evaluation of Skills"
  authors:
    - family-names: "Kevin"
      given-names: "Christopher"
    - family-names: "Raghavan"
      given-names: "Narendran"
    - family-names: "Puget"
      given-names: "Jean-Francois"
    - family-names: "Malani"
      given-names: "Roshni"
    - family-names: "Puvvadi"
      given-names: "Meghana"
    - family-names: "Abramovitch"
      given-names: "Moshe"
    - family-names: "Gupta"
      given-names: "Mohit"
    - family-names: "Akkiraju"
      given-names: "Rama"
    - family-names: "Prabhu"
      given-names: "Subodh"
    - family-names: "Dangi"
      given-names: "Yogesh"
    - family-names: "Luo"
      given-names: "Wei"
    - family-names: "Lee"
      given-names: "Seong Hee"
  year: 2026
  doi: "10.48550/arXiv.2608.20614"
  url: "https://arxiv.org/abs/2608.20614"