Radar check
What we found when we opened it. Every bar is one fact, with its date.
Last code pushed 13 Oct 2026, 14 days before this check.
Latest release 0.6.13 on 13 Oct 2026.
Public since February 2026, 8 months.
111,232 stars and 11,639 forks.
315 commits and 56 new issues in the last 90 days.
92 contributors. 116 pull requests merged in 90 days.
No security policy file says how to report a problem. No published security advisories.
How the bars are scored
- Alive. Days since the last code push: a week or less 100, a month 90, three months 70, six months 50, a year 30, eighteen months 15, older 0.
- Releases. Age of the latest GitHub release, on a similar scale; no bar when the project publishes no releases.
- Maturity. Age of the project: eight years 100, five 90, three 70, two 55, one 40, six months 25, younger 10.
- Adoption. Stars on a log scale: 100,000 stars is 100, 10,000 is 80, 1,000 is 60, 100 is 40.
- Activity. Commits in the last 90 days: 300 or more 100, 100 85, 30 65, 10 45, any 25.
- Community. People who have contributed: 100 or more 100, 30 80, 10 60, 5 40, 2 25, one 10.
- Security. A published security policy 80, none 40; published advisories are listed, not scored.
Checked by Repo Radar on 28 Oct 2026 from the repository on GitHub.
In simple English
This repository gives artificial intelligence coding assistants a strict set of rules to follow. Most coding helpers take the absolute easiest path and skip important tests or security checks. These files teach them how senior engineers build software step by step. You install them into your editor or agent tool of choice. It costs nothing in money, but you need to take time to set it up.
What it is
A collection of twenty-five structured workflows and nine command-line shortcuts for AI coding assistants. It covers the entire software development lifecycle from writing a specification to deploying code.
What it is good for
Solo builders who use AI tools like Claude Code or Cursor and want cleaner, safer code without having to constantly remind the assistant to write tests or follow security rules.
Summary (Key Points)from the README
Key features
- Lifecycle commands: Nine slash commands (`/spec`, `/plan`, `/build`, `/test`, `/constraints`, `/review`, `/webperf`, `/code-simplify`, `/ship`) map to each development phase.
- 25 skills: 24 lifecycle skills plus the `using-agent-skills` meta-skill, each with steps, verification gates, and anti-rationalization tables.
- Autonomous build mode: `/build auto` generates the plan and implements every task after one approved pass, committing each task individually and pausing on failures or risky steps.
- Five-axis code review: The review skill covers change sizing (~100 lines), severity labels (Nit/Optional/FYI), review speed norms, and splitting strategies.
- Test pyramid guidance: The testing skill sets a test pyramid of 80/15/5, plus red-green-refactor, test sizes, and DAMP over DRY.
- Interview-based requirements: `interview-me` asks one question at a time until about 95% confidence in what the user wants.
- Quality bar enforcement: `constraint-driven-development` writes CONSTRAINTS.md with default thresholds and catches agents silencing checks or skipping tests.
- Specialist personas: Four personas (code-reviewer, test-engineer, security-auditor, web-performance-auditor) support targeted reviews.
- Reference checklists: Seven checklists cover areas such as security, performance, accessibility, and observability.
Ways to install
| Platform or method | What it needs | Effort |
|---|---|---|
| Skills CLI (`npx skills add`) | Node.js with `npx`; installs into 70+ agents | Low |
| Claude Code marketplace | Claude Code; GitHub access (SSH keys or HTTPS workaround) | Low |
| Claude Code local plugin | Git clone; Claude Code with `--plugin-dir` | Low |
| Antigravity CLI | `agy` CLI; Git clone or repo URL | Low |
| Gemini CLI | `gemini` CLI; Git clone or repo URL | Low |
| Codex | Codex CLI v0.122+ | Low |
| Command Code | `cmd` CLI with built-in `skills` command | Low |
| Cursor | Manual copy into `.cursor/skills/` and `.cursor/rules/` | Medium |
| OpenCode | Manual copy into `.opencode/skills/` plus a project `AGENTS.md` | Medium |
| GitHub Copilot | Manual use of `agents/` and `.github/copilot-instructions.md` | Medium |
| Windsurf | Manual addition of skill contents to rules configuration | Medium |
| Kiro | Manual placement under `.kiro/skills/` | Medium |
Important
- Per-skill installs are incomplete: A single-skill `npx` install omits the repo-level `references/` directory, so links to shared checklists break; use a whole-repo install or clone the repository instead.
- Lifecycle commands are host-specific: The `/spec` through `/ship` commands exist in Claude Code, not in GitHub Copilot, which requires invoking skills by name.
- Antigravity wrapper limitation: In affected Antigravity CLI releases, legacy command wrappers are not discoverable, so the underlying namespaced skills must be invoked directly.
- Name collision in Claude Code: Where a built-in command shares a name such as `/review`, use the namespaced form `/agent-skills:review`.
- SSH clone failures: The Claude Code marketplace clones over SSH, so users without GitHub SSH keys must use the HTTPS URL or rewrite GitHub SSH URLs to HTTPS.
- Cursor requires manual sync: The repo does not ship a `.cursor/` folder, so users must copy skills and rules into their own project.
Written by an AI model from the project’s README; not part of the review, and it can be wrong.
- Enforces test-driven development and prevents your AI assistant from skipping tests.
- Works across seventy different agent tools and editors using a simple installation command.
- Encoded best practices help prevent common security mistakes and performance regressions.
- Requires manual configuration depending on which specific editor or agent tool you use.
- Some advanced features like shared reference files have minor portability gaps when installing only single skills.
- Using autonomous build loops requires close supervision so the agent does not go off track.
manually prompting your AI assistant with custom system instructions every time you start a new coding session
Claude Code
- obra/superpowers307k stars
Another popular collection of workflow skills and commands designed to guide AI coding assistants.
- mattpocock/skills305k stars
A curated set of workflow instructions and best practices tailored for AI coding tools.
- musistudio/claude-code-router38k stars
One local address that many coding assistants can talk to, pointed at whichever model the operator chooses.
- Kong/insomnia40k stars
A desktop application for sending requests to an interface and reading what comes back.
- microsoft/playwright-mcp39k stars
A server that lets an AI assistant drive a real browser, using the page's own structure rather than screenshots.
npx skills add addyosmani/agent-skills