# SkillEvaluator Documentation

## Instructions for AI Agents

- For clean Markdown of any page, append `.md` to the page URL
- For section-specific indexes, append `/llms.txt` to any section URL
- For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.nvidia.com/skills/skillevaluator/_mcp/server

## Docs

- [SkillEvaluator](https://docs.nvidia.com/skills/skillevaluator.md): An open-source, multi-tier framework for evaluating AI agent artifacts, starting with agent skills.
- [Quickstart](https://docs.nvidia.com/skills/skillevaluator/quickstart.md): Run a keyless quality check on a bundled skill, then add NVIDIA Build for your first model-backed evaluation.
- [Installation](https://docs.nvidia.com/skills/skillevaluator/installation.md): Install SkillEvaluator with uv, from source, or in Docker, and choose per-tier extras to keep the footprint small.
- [Providers & Credentials](https://docs.nvidia.com/skills/skillevaluator/configuration.md): Configure LLM providers, embeddings, and agent credentials for SkillEvaluator — and see exactly which commands run with no key at all.
- [Tier 1: Static and Security Validation](https://docs.nvidia.com/skills/skillevaluator/tier1-validation.md): Deterministic quality gates for skills — schema, quality, security, PII, license, code-integrity, Unicode, and script checks, plus optional LLM judging.
- [Tier 2: Semantic Deduplication](https://docs.nvidia.com/skills/skillevaluator/tier2-deduplication.md): Detect redundant content within one skill and overlapping skills across a collection using embedding similarity.
- [Tier 3: Live Agent Evaluation](https://docs.nvidia.com/skills/skillevaluator/tier3-live-evaluation.md): Run a real agent with and without your skill in a sandboxed environment, then measure Skill Lift across five dimensions.
- [Eval Datasets](https://docs.nvidia.com/skills/skillevaluator/eval-datasets.md): Author the evals/ directory that powers Tier 3 — the dataset format, generation with create-eval-dataset, the config.yml run policy, fixtures, and custom container environments.
- [Agents & Sandboxes](https://docs.nvidia.com/skills/skillevaluator/agents-and-sandboxes.md): Every agent Tier 3 can drive and every environment it can run them in — Docker tuning, local-mode sandbox policy, workspace modes, and readiness checks.
- [Custom Graders & Tasks](https://docs.nvidia.com/skills/skillevaluator/custom-graders.md): Bring your own grader (BYOG) or a complete native task (BYOT) when SkillEvaluator's default graders can't express what success means for your skill.
- [Reports & Results](https://docs.nvidia.com/skills/skillevaluator/reports.md): Every file SkillEvaluator writes — report formats, file naming, BENCHMARK.md, the Tier 3 results tree, and how to read dimensions, Skill Lift, and pass@k.
- [Performance Dashboard](https://docs.nvidia.com/skills/skillevaluator/performance-dashboard.md): Compare saved Tier 3 runs by skill, model, harness, score, tokens, time, and recorded cost in the optional local dashboard.
- [BENCHMARK.md Rollout and Backfill](https://docs.nvidia.com/skills/skillevaluator/benchmark-rollout.md): Regenerate published benchmark cards safely after scoring, renderer, or publication-policy changes.
- [Bring Skill Evaluation to Your Repository](https://docs.nvidia.com/skills/skillevaluator/repository-integration.md): Evaluate changed skills automatically with reusable GitHub and GitLab CI configurations.
- [GitHub Actions](https://docs.nvidia.com/skills/skillevaluator/github-actions.md): Bring SkillEvaluator to your GitHub repository with a reusable workflow, clear results and optional PR feedback.
- [Set Up CI with the CLI](https://docs.nvidia.com/skills/skillevaluator/cli-onboarding.md): Install SkillEvaluator, answer a few questions and bring automatic skill checks to your GitHub repository.
- [GitLab CI/CD](https://docs.nvidia.com/skills/skillevaluator/gitlab-ci.md): Add reusable SkillEvaluator Tier 1 checks to a GitLab repository.
- [Gate Your CI](https://docs.nvidia.com/skills/skillevaluator/ci-integration.md): Make SkillEvaluator a merge gate — exit codes, a copy-paste GitHub Actions workflow, collection validation, progressive adoption, and machine-readable results.
- [CLI Reference](https://docs.nvidia.com/skills/skillevaluator/cli-reference.md): Run tiers directly, explore expert subcommands, and look up SkillEvaluator flags and defaults.
- [Environment Variables](https://docs.nvidia.com/skills/skillevaluator/environment-variables.md): The complete reference of every environment variable SkillEvaluator reads — providers, embeddings, validation, and live evaluation.
- [Contributing](https://docs.nvidia.com/skills/skillevaluator/developer-guide.md): Set up a SkillEvaluator development environment, run the test suite, preview the docs locally, and submit a pull request that passes review.