design

archipelago: Fast‑track AI Agent Testing in RL Environments

Name: archipelago: Fast‑track AI Agent Testing in RL Environments
Author: Mercor-Intelligence

Run and benchmark your AI agents on reinforcement‑learning scenarios in minutes. Ideal for founders building intelligent products that need rapid validation.

From the repo · “Harness for running and evaluating AI agents against RL environments”

291 stars+6★ this week70 forksPythonGuide quality 9/10 · reviewed Oct 16Updated 10/23/2026100% free · open source

What it does

Archipelago allows you to evaluate and run AI agents against reinforcement learning environments, providing a harness for testing and comparing agent performance.

When to use it

•When you need to compare the performance of different AI agents in the same environment
•When you want to evaluate the effectiveness of an AI agent in a custom reinforcement learning environment
•When you need to standardize the evaluation process for AI agents across multiple environments

Ready-to-paste prompt

python examples/cartpole.py -c config.json -e CartPole-v1 -a PPO

Heads up: Ensure you have the correct version of Python and the required dependencies installed, as specified in the `setup.py` file, to avoid compatibility issues

Saves to your device

Use with Claude

New

Skip the builder — one click puts this in Claude, Cursor, Antigravity and more.

✅ Light setup

Installs with a command or two; your AI agent can do it for you.

Try it instantly — no install

Claude Code

mkdir -p ~/.claude/skills/archipelago && curl -fsSL https://workflowstacks.com/api/skills/archipelago/claude-skill -o ~/.claude/skills/archipelago/SKILL.md

Open in another AI app

Opens the app with this repo with the prompt ready to go — no copy-paste needed.

Connect the whole catalog (MCP)

claude mcp add --transport http workflowstacks https://workflowstacks.com/api/mcp

Adds a WorkflowStacks connector to Claude Code: search and load any skill here by chatting.

How archipelago: Fast‑track AI Agent Testing in RL Environments works

Codeflow

Free to inspect

Archipelago: Fast‑track AI Agent Testing in RL Environments is a very large Python project (~787k lines across 2837 code files, plus 98 test files). Setup is light: installs like a normal app. Reading the code is optional. Last commit this month, Apache-2.0 license, has a test suite.

Size

Very large codebase

~787k lines · 2837 code files · days to read — use, don't read

Setup

One-command install

Installs like a normal app. Reading the code is optional.

Runs on

Python

No API keys detected

Python 90%TypeScript 9%Shell 1%

Where to start reading

1
README.md
Start here — what it does and how to install it
2
agents/README.md
Inside agents/ — an example of what the AI is told to do
3
examples/hugging_face_task/README.md
A worked example — copy this to get going

What's in each folder

agents/Prompts, skills & agent definitions101 files

examples/Examples you can copy21 files

mcp_servers/documents/Folder401 files

grading/Folder396 files

mcp_servers/spreadsheets/Folder392 files

mcp_servers/pdfs/Folder390 files

mcp_servers/filesystem/Folder386 files

mcp_servers/mail/Folder345 files

READMEHas testsDocumentedNo CIExamples includedApache-2.0 licenseUpdated this month