Agent Skills
AI coding agents typically focus only on writing code and skip essential engineering practices like specs, tests, and reviews. Agent Skills is a framework designed to enforce these senior-engineer behaviors by embedding structured workflows into AI agent processes. It maps onto standard software development lifecycle phases and aims to make critical non-coding tasks mandatory, not optional.
- ▪AI coding agents default to producing code without creating specs, tests, or reviewable changes.
- ▪Agent Skills provides structured workflows with exit criteria to ensure AI follows reliable engineering practices.
- ▪The framework aligns with established SDLC phases used by companies like Google and Amazon.
- ▪Each 'skill' is a workflow, not reference material, designed to produce verifiable evidence at each step.
- ▪Agent Skills has gained over 26,000 GitHub stars, indicating broad interest in improving AI coding practices.
Opening excerpt (first ~120 words) tap to expand
Agent Skills May 3, 2026 A senior engineer’s job is mostly the parts that don’t show up in the diff. Specs. Tests. Reviews. Scope discipline. Refusing to ship what can’t be verified. AI coding agents skip those parts by default. Agent Skills is my attempt to make them not optional. The default behaviour of any AI coding agent is to take the shortest path to “done.” Ask for a feature and it writes the feature. It does not ask whether you have a spec, write a test before the implementation, consider whether the change crosses a trust boundary, or check what the PR will look like to a reviewer. It produces code, declares victory, and moves on. This is the same failure mode every senior engineer has spent their career learning to avoid.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at Addyosmani.
Discussion
0 commentsMore from Addyosmani
