About paretostudio-mcp
paretostudio-mcp connects MCP clients to a ParetoStudio library of prompts, agents, and skills. It supports searching and editing library entries, then installing agents and skills into Claude Code or Codex project layouts.
Key features
- 01List, search, and retrieve prompts, agents, and skills
- 02Create, update, and delete library entries
- 03Install agents and skills into project files
- 04Install an agent and its skills as a bundle
- 05Avoid overwriting existing files by default
Use cases
- 01Find a prompt or agent from an MCP client
- 02Create or revise entries in a ParetoStudio library
- 03Add an agent or skill to a Claude Code or Codex project
MCP support
It runs as an MCP server configured in clients such as Claude Code, Cursor, Claude Desktop, and Codex. It exposes tools for managing ParetoStudio prompts, agents, and skills, including installation into project files.
Configuration
Local server config is the same for every client. Copy it into your client’s MCP config file.
npx -y paretostudio-mcpPARETOSTUDIO_API_KEYSecretYour ParetoStudio API key (starts with psk_). Create one at https://paretostudio.io/settings under Developer.PARETOSTUDIO_API_URLAPI base URL. Only change this when pointing at a self-hosted or staging instance.Local servers need a runtime on your machine (e.g. Node.js 18+). Config file locations differ by client; see your client’s docs.
Tools
Local servers expose tools only once running on your machine, and the maintainer has not provided a tool list.
Details
From the project READMEparetostudio-mcp
Read your ParetoStudio prompt, agent, and skill library from Claude Code, Cursor, and any MCP client. Install agents and skills as files into Claude Code.
Quick Start
1. Get your API key
Go to paretostudio.io/settings and navigate to Developer, then create an API Key.
2. Configure your client
Claude Code (.mcp.json in your project root):
{
"mcpServers": {
"paretostudio": {
"command": "npx",
"args": ["-y", "paretostudio-mcp@latest"],
"env": {
"PARETOSTUDIO_API_KEY": "psk_your_key_here"
}
}
}
}
Cursor (Settings > MCP):
{
"mcpServers": {
"paretostudio": {
"command": "npx",
"args": ["-y", "paretostudio-mcp@latest"],
"env": {
"PARETOSTUDIO_API_KEY": "psk_your_key_here"
}
}
}
}
Claude Desktop (claude_desktop_config.json):
{
"mcpServers": {
"paretostudio": {
"command": "npx",
"args": ["-y", "paretostudio-mcp@latest"],
"env": {
"PARETOSTUDIO_API_KEY": "psk_your_key_here"
}
}
}
}
Codex (~/.codex/config.toml):
[mcp_servers.paretostudio]
command = "npx"
args = ["-y", "paretostudio-mcp@latest"]
[mcp_servers.paretostudio.env]
PARETOSTUDIO_API_KEY = "psk_your_key_here"
3. Use it
Ask Codex, Claude, or your MCP client to "list my prompts" or "install my agent" — the tools are available automatically.
Available Tools
| Tool | Description |
|---|---|
list_prompts | List all prompts in your library |
get_prompt | Get full prompt content (all 6 steps) |
search_prompts | Search prompts by title or description |
list_agents | List all agents |
get_agent | Get agent with full configuration |
search_agents | Search agents |
list_skills | List all skills |
get_skill | Get skill with full configuration |
search_skills | Search skills |
install_skill | Install skill for Claude or Codex project layouts |
install_agent | Download agent to .claude/agents/ |
install_bundle | Install agent + skills in one operation |
create_prompt | Create a new prompt |
update_prompt | Update a prompt |
delete_prompt | Delete a prompt |
create_skill | Create a new skill |
update_skill | Update a skill |
delete_skill | Delete a skill |
create_agent | Create a new agent |
update_agent | Update an agent |
delete_agent | Delete an agent |
Install tools never overwrite existing files by default. Pass overwrite: true
only when you intentionally want to replace an installed agent or skill.
For Codex project skills, call install_skill with client: "codex"; the skill
is written to .agents/skills/<name>/SKILL.md instead of .claude/skills/.
Write tools respect your plan's library caps (Free: 5 prompts / 2 agents / 2 skills; Pro and Team: unlimited).
Rate Limits
| Plan | Daily Requests | Write Access |
|---|---|---|
| Free | 50 | Yes — within plan caps (5 prompts / 2 agents / 2 skills) |
| Pro | 500 | Yes — unlimited |
| Team | 500 | Yes — unlimited, shared team library |
Environment Variables
| Variable | Required | Default | Description |
|---|---|---|---|
PARETOSTUDIO_API_KEY | Yes | -- | Your API key (starts with psk_) |
PARETOSTUDIO_API_URL | No | https://paretostudio.io | API base URL |
Source
The server source lives in the ParetoStudio application repository, because it
imports the same business rules the API enforces (src/rules/) — duplicating them
here would let the two drift apart. This repository is the public home of the
published package: the manifest registries read, the configuration examples, and
the documentation.
Releases are automatic: a version bump reaching main publishes to npm, and a CI
gate refuses a build whose server changed since the last bump.
Links
Privacy Policy
The server is a local process on your machine. It talks only to the
ParetoStudio API (https://paretostudio.io, or the PARETOSTUDIO_API_URL you
configure) using the API key you provide, and sends nothing anywhere else.
Privacy policy: paretostudio.io/privacy
Security
See SECURITY.md. Report vulnerabilities to [email protected] — not in public issues.
License
MIT — see LICENSE.
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.