Agentspace
Give your agent a persistent workspace: files, a code sandbox and a public URL to publish to.
- Transport
- Not stated
- Package
- —
- Registry id
- br.com.agentspace/agentspace
No install snippet on purpose. A working MCP config is a command, its arguments and an environment block — the last two are where API keys live, so this catalogue never stores them and cannot publish them. Follow the link above for the authors' own instructions.
It gives AI agents a persistent workspace where files outlive the session, with a sandbox to run code and a public address for publishing. It works with any agent, including Claude, ChatGPT, Codex, Cursor, GLM, and Gemini. It provides a folder that stays across conversations, days, and different agents.
It runs Python, Node, and shell in an isolated sandbox with the workspace folder mounted. It gives published content a public address that goes live immediately. It puts a person in charge who approves agents by name, can revoke their keys, and can review before anything goes live.
It lets agents in without a password, through a link the person approves and a key issued under the agent's name. It exposes the same tools over MCP, the REST API, and A2A on the same files.
Summary drafted from the project's own website. Every sentence is backed by text on that page and was reviewed before publishing.