Radar check
What we found when we opened it. Every bar is one fact, with its date.
Last code pushed 15 Apr 2026, 6 months before this check.
No releases published on GitHub; the code date above is the one to trust.
Public since February 2026, 8 months.
6 stars and 1 forks.
0 commits and 0 new issues in the last 90 days.
1 contributor. No pull requests merged in 90 days; changes go in directly. A small team carries it.
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 project provides pre-made instruction guides for AI coding tools like Claude and Cursor. It teaches your AI how to avoid deleting important databases or messing up git history. You download the rules using a simple command in your terminal. It is free to use under an open-source licence. You only spend time setting up the rules you actually need.
What it is
A repository of prompt instructions and plugins organised by audience and task for AI coding assistants. It covers safety guards, git workflows, multi-agent setups, security reviews, and model context protocol servers.
What it is good for
Solo builders who want their AI coding assistant to stop making common mistakes like force pushing to git or running dangerous database commands without backups.
Summary (Key Points)from the README
Key features
- Broad catalog: 94 skills and 7 plugins, organized by audience across safety, developer, MLOps, RAG, security, and 2026 suites.
- Cross-platform reach: The skills CLI works with Claude Code, Cowork, Cursor, Codex, Windsurf, and 35+ other agents that support skills.sh.
- Automatic activation: Skills trigger from what you are doing, so no special commands are needed.
- MCP integrations: Plugins ship pre-configured MCP connections to services such as Datadog, PagerDuty, Jira, and Pinecone, all marked optional.
- Slash commands: The plugins add 20+ commands, such as `/incident-response`, `/postmortem`, and `/security-scan`.
- Cost reduction claim: The prompt-caching-patterns skill claims cost reductions of 73%+ through semantic caching.
- Claude 4.6 and Agent SDK coverage: Dedicated suites cover the 1M context window, extended thinking, the Agent SDK, and remote MCP OAuth.
- Evaluation skills: The evals suite covers LLM-as-judge, regression testing, and golden sets of 20-100 must-not-regress cases.
Ways to install
| Platform or method | What it needs | Effort |
|---|---|---|
| Skills CLI (`npx skills add`) | npx and a supported AI agent such as Cursor, Codex, or Windsurf | One command |
| Claude Code / Cowork marketplace | Claude Code CLI or the Cowork desktop app, using `/plugin` commands | One marketplace add, then one command per plugin |
| Direct plugin install | Claude Code or Cowork, plus the plugin path in the repository | One command per plugin |
Important
- API keys: MCP plugins need your own keys in the plugin's `.mcp.json`. The README says never to commit them and to use environment variables.
- Production credentials: The README warns to be careful when connecting production credentials to MCP servers.
- Enable only what you need: Every MCP integration is optional, but review `.mcp.json` files before copying them.
- Confirmation not guaranteed: Skills ask for confirmation before destructive operations, yet the README notes that AI agents may interpret skills differently.
- No warranty: Skills are provided "as is," and the authors accept no liability for damages or data loss. Keep backups.
- Dated pricing: The token-cost-analyzer skill uses "2026 pricing," so check current rates before relying on its figures.
- External plugin docs: Full plugin documentation is in the official Claude Code plugins guide, not in this README.
Written by an AI model from the project’s README; not part of the review, and it can be wrong.
- Works with dozens of popular AI coding tools and assistants.
- Organised cleanly by role so you only install what you need.
- Includes safety guardrails that prevent accidental data loss.
- Requires familiarity with command-line tools to install and manage.
- Some advanced plugins need external API keys and extra configuration.
- AI assistants may occasionally ignore instructions depending on the model.
manually writing custom system prompts and rules for every new AI coding project
Cursor
- anthropics/claude-code154k stars
An official command-line interface for running AI coding agents with local file access.
- Aider-AI/aider50k stars
A command-line tool that lets you pair program with AI directly in your local git repository.
- 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 latestaiagents/agent-skills --all