[GitHub Trending] NVIDIA/SkillEvaluator
8.5 relevance
Score Breakdown
technical depth 9
novelty 9
actionability 8
community 7
strategic 8
personal 9
Scored daily by a customisable AI persona to surface the most relevant engineering leadership news.
NVIDIA's multi-tier AI agent skill evaluation framework with quality gates, cutting-edge for agent quality.
Summary
NVIDIA released SkillEvaluator, an open-source CLI tool that validates AI agent skills and plugins using 500+ static rules without requiring an LLM API key. It runs three tiers of checks: static quality and security (Tier 1), deduplication and compatibility (Tier 2), and live agent evaluation measuring skill lift (Tier 3). The tool integrates with CI pipelines on GitHub and GitLab, supports agents like Claude Code and Codex, and includes a harness check that tests whether a skill loads across six different agent runtimes.