Agent Skills Hub

codex

Delegates coding tasks to the OpenAI Codex CLI for features, refactoring, PR reviews, and automated issue fixing within Git repositories.

release

Automates the release workflow for Worktrunk, including version bumping, CHANGELOG generation, contributor crediting, and publishing to crates.io and GitHub.

context-degradation

Diagnose, isolate, and mitigate LLM context failures like lost-in-middle, poisoning, distraction, and context clash to improve agent reliability.

gh-local

Run GitHub Actions CI workflows locally using nektos/act in Docker. Test your CI configurations, debug workflow failures, and validate pipeline changes without pushing code to GitHub.

skill-architect

Expert consultant for designing and building high-quality, consistent AI agent skills. Guides you through discovery, architecture, and creation phases to ensure reliable, composable, and efficient skill delivery.

component-workflow

Guidance for creating and integrating new ECS components, including class design, JSON serialization, custom editor UI implementation, and DI registration.

scholar-evaluation

Systematically evaluate scholarly work using the ScholarEval framework, providing structured, quantitative, and qualitative assessment across research quality dimensions with actionable feedback.

fumadocs-i18n

Automate complete internationalization for Fumadocs. Configures multi-language routing, language switcher, sidebar filtering, and content structure for Next.js projects.

project-overview

Complete project architecture and structure guide for LobeHub. Use for codebase exploration, project organization, file location, and architectural context.

paper-figure

Generate publication-quality figures, charts, and LaTeX tables from experiment data for academic papers.

writing-skills

A Test-Driven Development (TDD) framework for writing agent skills, using pressure scenarios to ensure documentation guides agent behavior effectively.

recursive-decomposition

Handles large-scale tasks by automatically breaking them down into manageable, recursive sub-tasks to overcome context window limits and improve reasoning accuracy on large codebases and document sets.