rapidreview-io

Merv — AI skill for Claude Code

AI community

Merv — agent-run ML research as gated, reviewable experiment workflows.

How to install Merv

This entry records only its repository, not the path inside it, so there is no exact command to give. Open rapidreview-io/Merv and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Merv does

Merv — agent-run ML research as gated, reviewable experiment workflows. Plugin for Claude Code, Codex, Cursor, Gemini CLI, and other MCP clients.

Alternatives in AI

AIHawk — Open-source AI browser agent: describe any task in plain language and it autonomously browses, clicks, types a 30.3k ★
Firecrawl MCP Server — 🔥 Official Firecrawl MCP Server - Adds powerful web scraping and search to Cursor, Claude and any other LLM c 6.1k ★
Skills Manage — Desktop app to manage AI coding agent skills across Claude Code, Cursor, Gemini CLI, Codex, and 20+ platforms 2.2k ★

README

Merv

Merv is a plugin for agentic coding platforms that helps agents run machine learning research as gated, reviewable experiment workflows.

Documentation: [rapidreview.io/docs/merv](https://rapidreview.io/docs/merv)

It is designed to work with Claude Code, Codex, GitHub Copilot CLI, Cursor, Gemini CLI, Qwen Code, Kilo Code, Hermes Agent, OpenCode, OpenHands, Replit Agent, and other MCP-capable agent platforms. It includes a frontend for humans to observe agent behavior ranging from macro research strategy to experiment execution specifics.

The goal is to give research agents enough structure to plan experiments, execute them, review results, and reflect on the project direction to handle open-ended research problems.

Experiment-level workflow

Experiment workflow: Plan, Design review, Execute, Results review, Complete. Rejected reviews send work back to Execute or Plan.

Each experiment begins with a generated plan that is adversarially reviewed by another agent. The plan/review loop persists until the reviewer approves the plan. After approval, the agent proceeds to execution. When it is done, it submits a report that is adversarially reviewed by a different agent. The reviewer can send the agent back to execution to fix something in the execution or the report, or it can send it back to the planning stage if the experiment proved faulty.