skill-up vs SkillOpt
Agent skill evaluation and optimization compared
skill-up tests agent skills, agents, and workspaces with repeatable cases and structured reports. SkillOpt optimizes a Markdown skill document using scored task trajectories and validation, while keeping the target model’s weights unchanged.
skill-up: Evaluate and Improve Agent Skills
skill-up is a Go CLI for testing Agent Skills, agents, and workspaces with repeatable cases and structured reports. It suits teams that want to catch behavior regressions or iteratively improve Skills using evaluation results.
SkillOpt: Train Reusable Skills for Frozen LLM Agents
SkillOpt improves natural-language agent skills using scored task trajectories and validation-gated edits, without changing the target model's weights. It is aimed at teams that can evaluate repeatable tasks and want to deploy a reusable skill document.
| skill-up | SkillOpt | |
|---|---|---|
| Language | Go | Python |
| License | Apache-2.0 | MIT |
| Stars | 1.1k | 18k |
| Forks | 95 | 1.7k |
| Last analyzed | Oct 7, 2026 | Oct 3, 2026 |
Key differences
- skill-up is a Go CLI for evaluating skills, agents, and workspace tasks; SkillOpt focuses on improving a reusable skill document for repeatable tasks.
- skill-up runs configured test cases and grades results with checks, scripts, or an agent judge; SkillOpt proposes bounded text edits from scored trajectories and keeps updates that pass validation.
- skill-up reports results in formats including JSON, JUnit XML, and HTML for local review or CI; SkillOpt can produce a deployable `best_skill.md` and offers an optional Gradio WebUI.
- skill-up supports built-in engines including Claude Code, Codex, and Qoder CLI, plus custom engines; SkillOpt supports chat and execution backends, benchmark adapters, and an offline SkillOpt-Sleep workflow.
- skill-up is licensed under Apache-2.0 and lists 1.1k stars; SkillOpt is MIT-licensed and lists 18k stars.
- skill-up requires Go 1.25 to build from source and lists Node.js as a runtime requirement; SkillOpt's optimization process requires a separate model to propose edits and useful evaluation and held-out validation data.
Choose skill-up if you…
- need repeatable evaluations of skills, agents, or workspace tasks with reports for CI.
- want to compare Skill-enabled runs with runs that omit the Skill.
- prefer a Go CLI with built-in agent engines and configurable evaluation cases.
Choose SkillOpt if you…
- want to optimize a reusable Markdown skill while keeping the target model's weights frozen.
- have scored task trajectories and held-out validation data for judging candidate edits.
- need training and evaluation tooling, benchmark adapters, or an offline workflow for consolidating validated skills.
This comparison is generated with AI from the OSRepos analyses of both projects. Always check each project's repository and documentation before choosing.