Our take
Activeregular commits and releases- Activity
- 77
- Community
- 71
- Issues
- 67
- Pull requests
- 48
+10,572 stars in the last 30 days
- Merges more pull requests than 81% of the projects we track
- Slower to merge pull requests than 83% of the projects we track
This is a broad, actively maintained set of engineering workflows, best suited to teams willing to adopt its verification-first process and adapt installation details to their agent.
Good fit if
- You want AI agents to follow explicit workflows for planning, implementation, testing, review, and release.
- Your team uses one of the documented agent integrations or can install plain Markdown skills into its tool.
- You value a project with recent commits and releases, contributions from many developers, and issues typically handled quickly.
Look elsewhere if
- You need a single, turnkey integration that works identically across every agent. Setup and command behavior differ by host.
- You want a software library or runtime component rather than instruction files and workflow guidance.
- You need every skill to be independently installable with all shared reference checklists included. The README documents a portability gap for per-skill installs.
All health signals
| Last commit | 2026-10-13 (6 days ago) |
|---|---|
| Commits, last 90 days | 100+ |
| Releases, last 12 months | 15 (latest 0.6.13, 2026-10-13) |
| Contributors | 81 (top contributor: 52% of commits) |
| Issues closed, last 90 days | 11+ (typically closed in 2 days) |
| Pull requests merged, last 90 days | 69 (typically merged in 3 days) |
| Project age | 8 months |
Checked on 2026-10-20 with the GitHub API.
Project history
Daily GitHub snapshots since Oct 7, 2026. History covers stars and forks; health signals are not tracked historically.
Stars
117,999Forks
12,345Change since Nov 15, 2026: +382 stars·+42 forks
Overview
addyosmani/agent-skills is a pack of portable Markdown workflows for AI coding agents. It aims to prevent agents from taking shortcuts by encoding engineering practices, verification gates, and quality standards across the development lifecycle.
The workflows cover defining requirements, planning, building, testing, reviewing, and shipping. They can be installed into multiple agent tools, so teams can apply a consistent process across different supported platforms rather than relying on one agent-specific prompt.
Key Features
- Provides 25 skills, including workflows for specification, task planning, incremental implementation, testing, security, and deployment.
- Offers nine lifecycle slash commands for Claude Code, from
/specand/buildto/reviewand/ship. - Uses step-by-step processes, exit criteria, and evidence requirements to make verification explicit.
- Includes anti-rationalization tables that address common reasons an agent might skip a workflow step.
- Packages four specialist agent personas and seven reference checklists for reviews and recurring quality checks.
- Supports installation or setup for tools including Claude Code, Cursor, Codex, Copilot, Gemini CLI, and Antigravity CLI.
Use Cases
- A developer starting a feature can use the spec and planning skills to turn an underspecified request into verifiable implementation tasks.
- A team adopting AI coding agents in an existing codebase can use the adoption guide's incremental, verification-first approach.
- Engineers can apply testing, security, and code-review workflows when asking an agent to change behavior or prepare a change for merge.
- Teams using more than one coding agent can reuse the shared skills to establish similar working practices across supported tools.
What you need
Detected in the repository
- JavaScript (main language on GitHub)
- Automated checks on GitHub Actions
License in plain words
MITpermissive
- Commercial use: yes
- Modify and redistribute: yes
- You must keep: the copyright and license notice
- Share your changes: no
A summary, not legal advice: the LICENSE file is what applies.
Getting Started
Install the full pack with the skills CLI:
npx skills add addyosmani/agent-skills
For host-specific setup, individual skills, and adoption guidance, see the README.
Alternatives
- compound-engineering-plugin: Packages skills as a plugin with brainstorming, planning, implementation, review, and learning workflows for Claude Code, Cursor, and Codex.
- agent-skills: Emphasizes specification writing, review, testing, and autonomous sprint workflows with explicit quality gates.
- agentic-engineering: Focuses on documentation-first contracts, project memory, decisions, and Git workflow, with merges left to humans.
- agent-feed: Installs a repository-owned workflow with task and review gates, checks, and tool adapters rather than offering only a portable skills pack.
| Project | Language | License | Stars | Status |
|---|---|---|---|---|
| agent-skills | JavaScript | MIT | 109k | Active |
| compound-engineering-plugin | TypeScript | MIT | 25.4k | Active |
| agent-skills | Shell | MIT | 2 | Active |
| agentic-engineering | – | MIT | 5 | Quiet |
| agent-feed | Python | MIT | 4 | Quiet |
Considerations
The project is actively maintained, with regular commits and releases, and its contributor base is not concentrated around one person. It is still a relatively young project, so teams should validate that its workflows and integration instructions fit their own practices and agent versions. The README also notes that installing a single skill through the skills CLI does not copy the repository-level references/ directory, and that some Antigravity CLI releases have a limitation with legacy wrapper commands. The repository provides CI and evaluation case files, but the supplied requirements data does not report a test suite.
Found this useful?
Share it with someone who would like agent-skills.