Autoresearch Agent plugin: what it installs and how to add it
Runs measurable experiments to optimise files, using built-in evaluators to compare changes against a chosen metric.
From the Marketplace: claude-code-skills marketplace
Install
> /plugin marketplace add alirezarezvani/claude-skills> /plugin install autoresearch-agent@claude-code-skillsRun both lines inside Claude Code: the first adds the marketplace, the second installs the plugin.
What is inside
- 6skills
- 1agent
About this plugin
The plugin sets up and runs experiment loops that make changes to a file, evaluate each result against a measurable metric and track the outcome. It includes eight built-in evaluators and supports configurable loop intervals from 10 minutes to monthly.
Use it when you want to optimise a file through repeated, metric-driven experiments and need to check or resume a run. The plugin installs six skills for setup, running, looping, status checks and resuming, plus an experiment-runner agent.
What's inside
- 6 skills:
ar-resume,ar-status,autoresearch-agent,loop,runandsetup - 1 agent:
experiment-runner
More from claude-code-skills
- A11y AuditScans and fixes WCAG 2.2 accessibility issues in React, Vue, Angular, Svelte, Next.js and HTML projects.Marketplace: claude-code-skillsGitHub stars: 27.2k
- AEOAudits and optimises content for citation by AI answer engines including ChatGPT, Perplexity, Claude, Gemini and Mistral.Marketplace: claude-code-skillsGitHub stars: 27.2k
- Agent HarnessRuns bounded agentic loops that turn goals into verified tasks and close only when checks pass or are explicitly waived.Marketplace: claude-code-skillsGitHub stars: 27.2k
- Agent Launcher SkillsBuilds, launches, grades, and schedules Claude Managed Agents in your Anthropic account through bounded workflows and loops.Marketplace: claude-code-skillsGitHub stars: 27.2k
- Agent MemoryAdds tiered memory for Claude Code, recalling relevant project context while requiring repeated evidence and human approval before promotion.Marketplace: claude-code-skillsGitHub stars: 27.2k
- AgentHubRuns parallel Claude Code agents in isolated worktrees, evaluates their competing solutions and merges a selected winner.Marketplace: claude-code-skillsGitHub stars: 27.2k
Related plugins
- SuperpowersOfficialGuides Claude Code through planning, TDD, debugging, code review, and subagent-driven software development.Marketplace: claude-plugins-officialGitHub stars: 294k
- Matt Pocock SkillsOfficialAdds reusable engineering skills for specification, TDD, code review, debugging, domain modelling and project handoffs.Marketplace: claude-plugins-officialGitHub stars: 274k
- Understand AnythingAnalyses codebases and documents into an interactive knowledge graph for exploring files, symbols, dependencies, and project structure.Marketplace: claude-communityGitHub stars: 85k
- Chrome DevTools MCPOfficialControls and inspects a live Chrome browser for debugging, performance analysis, network checks, and Puppeteer automation.Marketplace: claude-plugins-officialGitHub stars: 52.9k
Skills, not plugins
Prefer a single skill for development work? 471 skills in the same category, each installed on its own.