Agent Plugins Marketplace
← All plugins

auto-loop

v4.2.1

Autonomous code improvement loop, Claude Code edition - autocode: hypothesis-driven parallel experiments with in-session experimenters routed by the plugin agents, a live experiment board published as an Artifact, unlazy-backed termination gates, and kept changes collected as one squash commit each on a worktree branch that ends in a PR.

Claude Code1 Skill

By weedmoLicense: MIT1 GitHub starsUpdated 3 days ago

Directory evidence

Runtimes
Claude Code
Parsed components
1 skill or MCP entry
Source updated
Sep 21, 2026
Manifest status
Canonical path parsed

The directory validates manifest shape and source location. It does not execute the plugin or provide a security endorsement. Review the indexing methodology →

Install auto-loop for Claude Code

Installs for the current user
claude plugin marketplace add IchenDEV/agent-plugin-mkt
claude plugin marketplace update agent-plugin-marketplace
claude plugin install auto-loop-2@agent-plugin-marketplace

Paste and run these commands in a terminal with Claude Code. They add and refresh the PluginsMP catalog, then install this plugin.

The installer fetches third-party code from the source repository shown on this page. This directory validates manifest structure and source location, but does not perform a security audit; review the manifest, components, and source before installing.

Get the source manually
git clone https://github.com/weedmo/skills

Clone the source repository, then follow its setup instructions to add the plugin to a compatible client. The plugin root is plugins/auto-loop-claude/.

Plugin files

plugins/auto-loop-claude/
├── .claude-plugin/plugin.json
└── skills/autocode/SKILL.md

Included Skills1

autocodeskills/autocode/SKILL.md

Hypothesis-driven parallel code improvement loop. A strategist on the expensive tier proposes falsifiable hypotheses with pre-committed if_confirmed / if_refuted actions; experimenters routed by difficulty implement them concurrently in git worktrees; the coordinator measures serially, keeps by arithmetic, and feeds every result back to the strategist so the frontier is revised as results arrive. Kept changes land as one squash commit each on an experiment branch in its own worktree — the user's checkout never moves — and the run ends by opening a PR of the kept changes (never merged). Subcommands: init [N] [--spec <path>], run [--parallel N] [--pr <base> | --no-pr], status, resume.

Plugin manifests1

plugins/auto-loop-claude/.claude-plugin/plugin.json
{
  "name": "auto-loop",
  "version": "4.2.1",
  "description": "Autonomous code improvement loop, Claude Code edition - autocode: hypothesis-driven parallel experiments with in-session experimenters routed by the plugin agents, a live experiment board published as an Artifact, unlazy-backed termination gates, and kept changes collected as one squash commit each on a worktree branch that ends in a PR.",
  "author": {
    "name": "weedmo"
  },
  "repository": "https://github.com/weedmo/skills",
  "homepage": "https://github.com/weedmo/skills",
  "license": "MIT",
  "keywords": [
    "loop",
    "autonomous",
    "autocode",
    "performance",
    "hypothesis"
  ],
  "skills": "./skills/"
}

For maintainers

If you maintain this plugin, link to this source-backed listing from your README so users can review its manifest and indexed components.

[auto-loop on Agent Plugins Marketplace](https://pluginsmp.com/plugins/auto-loop-2)