Analysis updated 2026-05-18
Guide an AI coding assistant through writing a detailed spec before implementing a feature.
Ensure an AI assistant runs security checks and tests before proposing code changes.
Set up automated code review workflows where the AI follows a checklist of verification steps.
Train an AI tool to break large tasks into small, testable increments instead of monolithic changes.
| addyosmani/agent-skills | michaelcade/90daysofdevops | vscodium/vscodium | |
|---|---|---|---|
| Stars | 30,187 | 29,449 | 31,265 |
| Language | Shell | Shell | Shell |
| Setup difficulty | easy | easy | easy |
| Complexity | 1/5 | 3/5 | 2/5 |
| Audience | developer | ops devops | developer |
Figures from each repo's GitHub metadata at analysis time.
Agent Skills is a collection of structured workflow guides designed to be fed to AI coding assistants, programs that help developers write, test, and ship software. The problem it addresses is that AI coding tools, left to their own defaults, often skip the steps a careful human engineer would follow: writing a spec before touching code, breaking work into small testable pieces, running security checks, or doing a proper review before merging a change. Each "skill" in the collection is a Markdown document (a plain text file with simple formatting) that spells out a specific workflow with clear steps, verification checkpoints, and common anti-patterns to avoid. There are 20 skills total, organized around the phases of building software: defining what to build, planning the work, writing the code, verifying it works, reviewing it, and shipping it. Seven slash commands (short instructions like "/spec" or "/test") activate the right skills automatically at each phase. You would use this when you are working with an AI coding assistant and want it to behave more like a disciplined senior engineer rather than jumping straight to writing code. The skills install into supported tools by adding files to specific directories or running a plugin command. The repository itself is written in Shell and the skills are plain Markdown, so they work with any AI agent that accepts system prompt files or instruction documents.
A collection of 20 structured workflow guides for AI coding assistants that enforce disciplined engineering practices, specs, testing, security checks, and code review, instead of jumping straight to coding.
Mainly Shell. The stack also includes Shell, Markdown.
Use freely for any purpose including commercial, as long as you keep the copyright notice.
Setup difficulty is rated easy, with roughly 5min to a first successful run.
Mainly developer.
This repo across BitVibe Labs
Verify against the repo before relying on details.