Creates custom auto-improve loops for any feature in any codebase — analyzes features, builds evaluation rubrics, and generates autonomous iterative improvement skills
npx claudepluginhub aminry/auto-claude-plugins --plugin auto-improveA marketplace of Claude Code plugins for automation, quality, and productivity. These plugins create autonomous iterative loops that analyze your codebase, interview you once, then run hands-off — making changes, evaluating results, and keeping only what works.
# Add the marketplace
/plugin marketplace add aminry/auto-claude-plugins
# List available plugins
/plugin list
# Install a plugin
/plugin install auto-improve@auto-claude-plugins
/plugin install auto-develop@auto-claude-plugins
| Plugin | Description | Version |
|---|---|---|
| auto-improve | Creates custom auto-improve loops for any feature — builds evaluation rubrics and generates autonomous iterative improvement skills | 1.0.1 |
| auto-develop | Creates custom auto-develop loops for any feature — builds capability specs and generates autonomous iterative development skills that add new functionality | 1.0.0 |
Improves the quality of existing features. Builds an evaluation rubric (40-100 yes/no questions), then runs a loop that scores output, identifies the highest-impact failure cluster, makes a single fix, and keeps only changes that improve scores.
/auto-improve:create <feature> # Build rubric + improvement loop
/auto-improve:evolve-rubric <feature> # Expand rubric after running cycles
Initialize → Generate & Evaluate → Analyze Failures → Implement Fix → Compare & Decide → Loop
Adds new functionality to existing features. Builds a capability specification (10-20 tiered capabilities with dependencies and acceptance criteria), then runs a loop that selects the next eligible capability, plans the implementation, writes code, verifies acceptance criteria, and runs regression checks.
/auto-develop:create <feature> # Build spec + development loop
/auto-develop:evolve-spec <feature> # Expand spec after running cycles
Initialize → Select Capability → Plan → Implement → Verify → Regression Check → Accept/Reject → Loop
| Situation | Plugin |
|---|---|
| Feature exists but output quality is poor | auto-improve |
| Feature exists but is missing functionality | auto-develop |
| Want to hit a quality score target | auto-improve |
| Want to build out a feature roadmap | auto-develop |
| Fixing bugs, formatting, accuracy | auto-improve |
| Adding endpoints, options, capabilities | auto-develop |
Both plugins auto-detect your tech stack (Node.js, Python, Go, Rust, Ruby, Java/Kotlin) and app type (web backend, frontend, CLI, data pipeline, ML/AI, library, mobile).
Add to your project's .claude/settings.json to auto-enable for all team members:
{
"extraKnownMarketplaces": {
"auto-claude-plugins": {
"source": {
"source": "github",
"repo": "aminry/auto-claude-plugins"
}
}
},
"enabledPlugins": {
"auto-improve@auto-claude-plugins": true,
"auto-develop@auto-claude-plugins": true
}
}
auto-claude-plugins/
├── .claude-plugin/
│ └── marketplace.json ← Marketplace catalog
├── plugins/
│ ├── README.md ← Plugin development guide
│ ├── auto-improve/ ← Auto-improve factory plugin
│ │ ├── .claude-plugin/
│ │ │ └── plugin.json
│ │ ├── commands/
│ │ │ ├── create.md
│ │ │ └── evolve-rubric.md
│ │ └── README.md
│ └── auto-develop/ ← Auto-develop factory plugin
│ ├── .claude-plugin/
│ │ └── plugin.json
│ ├── commands/
│ │ ├── create.md
│ │ └── evolve-spec.md
│ └── README.md
├── README.md
└── LICENSE
These plugins were inspired by autoresearch by Andrej Karpathy — the idea of using automated iterative loops with evaluation functions to continuously improve and expand software.
See Plugin Development Guide for instructions on adding new plugins.
MIT
Tools to maintain and improve CLAUDE.md files - audit quality, capture session learnings, and keep project memory current.
Browser automation and end-to-end testing MCP server by Microsoft. Enables Claude to interact with web pages, take screenshots, fill forms, click elements, and perform automated browser testing workflows.
20 modular skills for idiomatic Go — each under 225 lines, backed by 48 reference files, 8 automation scripts (all with --json, --limit, --force), and 4 asset templates. Covers error handling, naming, testing, concurrency, interfaces, generics, documentation, logging, performance, and more. Activates automatically with progressive disclosure and conditional cross-references.
Testing, security, and code quality experts - code review, penetration testing, QA automation
Comprehensive PR review agents specializing in comments, tests, error handling, type design, code quality, and code simplification
Real-time statusline HUD for Claude Code - context health, tool activity, agent tracking, and todo progress