antonbabenko

antonbabenko / agent-plugins

Public

A collection of agent plugins for improving productivity, automating workflows, and making AI coding agents work better together.

agent-plugins agent-skills claude-code codex-skills skills
15
0
89% credibility
Found May 21, 2026 at 15 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
AI Summary

Agent Plugins is a marketplace of add-ons for AI coding assistants like Claude Code, Cursor, Copilot, and Codex. It offers three plugins: code-intelligence (helps the assistant use the right tools and be honest about its work), terraform-skill (helps catch infrastructure mistakes), and claude-delegator (gives the assistant expert backup for tough decisions). The plugins are designed to make AI assistants more reliable and trustworthy by teaching them to double-check their work and admit when they need help.

How It Works

1
🤖 You start using an AI coding assistant

You try Claude Code, Cursor, or another AI assistant to help write code faster.

2
😕 The assistant makes frustrating mistakes

It sometimes uses the wrong search tool, misses references, or doesn't catch problems before they happen.

3
✨ You discover Agent Plugins

A collection of add-ons that teach your AI assistant to work more carefully and honestly.

4
You pick what you need
🔍
code-intelligence

Makes the assistant double-check its work and tell you when it switches tools

🏗️
terraform-skill

Helps the assistant catch infrastructure mistakes before they break things

🤝
claude-delegator

Gives your assistant expert backup when facing tricky decisions

5
⚙️ You install the plugins in one line

A simple command adds the plugins to your assistant so it's smarter from now on.

6
🎯 Your assistant works more carefully

It picks the right tools, admits when it's unsure, and catches problems early.

🚀 You ship better code with confidence

Your AI assistant now thinks before it acts, so your projects are more reliable.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 15 to 15 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is agent-plugins?

Agent-plugins is a plugin marketplace for AI coding agents that bundles three productivity plugins: code-intelligence, terraform-skill, and claude-delegator. It works across Claude Code, Cursor, Copilot, Gemini CLI, OpenCode, and Codex. The code-intelligence plugin helps agents pick the right tool for the task (language server vs text search vs fuzzy search) and discloses when it has to swap tools. Terraform-skill routes Terraform requests to handle failure modes like identity churn, secret exposure, and state corruption before generating code. Claude-delegator gives the agent GPT and Gemini expert subagents for second opinions on architecture, security, and planning decisions. Installation uses standard CLI commands like `npx skills add` or `/plugin install`.

Why is it gaining traction?

The hook is honest-by-construction tooling. Instead of trusting an AI that silently swaps search tools or skips validation steps, this plugin makes the agent disclose exactly what it is doing on the first line. For Terraform specifically, it prevents the common trap of generating plausible-looking HCL that breaks on apply. The multi-agent consensus feature (GPT + Gemini + Claude iterating until all three agree) solves a real pain point for high-stakes architecture decisions. Cross-platform compatibility means developers can use the same discipline regardless of which AI assistant they prefer.

Who should use this?

Infrastructure engineers working with Terraform who are tired of AI-generated code that fails during apply. Developers using Claude Code or Codex who want better tool selection transparency. Teams making architecture decisions that need a second opinion from multiple AI models before committing. It is less useful for general-purpose coding without Terraform involvement.

Verdict

The credibility score of 0.8999999761581421% reflects a very young project with only 15 stars, so treat this as early-stage software. The documentation is thorough and the author (Anton Babenko) has Terraform credibility, which helps. Try the code-intelligence plugin first if you use Claude Code, but hold off on betting critical workflows on terraform-skill until it has more community battle-testing.

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.