Is Agent Collab safe to use?

Agent Collab has 1 known CVE tracked by MCPpedia. You can verify these on OSV.dev by searching for "@openai/codex". Review the Security section above for details before installing.

How do I install Agent Collab?

Agent Collab supports copy-paste install configs on its MCPpedia page for Claude Desktop, Cursor, and Claude Code. Scroll to the Quick Install section and select your client.

What AI clients work with Agent Collab?

Agent Collab is compatible with claude-desktop, cursor, claude-code. It uses stdio and sse transport.

Is Agent Collab actively maintained?

Agent Collab is actively maintained — last commit was 8 days ago. It has 32 GitHub stars.

Agent Collab

Name: Agent Collab
Author: zebbern

@openai/codex · by zebbern

Run Codex and Cursor as background workers inside Claude Code. Reviews, tasks, and rescues with tracked jobs, live progress, and session handoffs.

32 16.2M/wk GitHub npm

1 open CVE

Apache-2.0 license

Actively maintained

Last commit 8d ago

Works with most clients

Transport: stdio, sse

0 tools

Grade F

MCPpedia last refreshed this data 1d ago

Edit this pageView history

Agent Collab is an MCP server that run Codex and Cursor as background workers inside Claude Code. Reviews, tasks, and rescues with tracked jobs, live progress, and session handoffs. Its tool list has not been published yet over stdio and sse, requires no API key, and scores 57/100 on MCPpedia's security, maintenance and efficiency rubric.

AI / ML

Step 1

Install in your client

Config is the same across clients — only the file and path differ.

Supported in Claude Desktopstdio, sse · Node 18+

Paste into ~/Library/Application Support/Claude/claude_desktop_config.json

{
  "mcpServers": {
    "agent-collab": {
      "args": [
        "-y",
        "@openai/codex"
      ],
      "command": "npx"
    }
  }
}

Are you the author?

Add this badge to your README to show your security score and help users find safe servers.

Embed in your READMEAbout badges →

[![MCPpedia Score](https://mcppedia.org/api/badge/agent-collab)](https://mcppedia.org/s/agent-collab)

Read me

What Agent Collab does

Delegate work from Claude Code to other AI coding agents.

Test This Server

Run this in your terminal to verify the server starts. Then let us know if it worked — your result helps other developers.

npx -y '@openai/codex' 2>&1 | head -1 && echo "✓ Server started successfully"

After testing, let us know if it worked:

Loading README…

Scored, not listed

Why this score

Five weighted categories — click any category to see the underlying evidence.

Score breakdown

57/100across 5 weighted dimensions

How we score →

0255075100

−43

Security

Maintenance

Efficiency

Documentation

Compatibility

Categoriesclick a row to see evidence

Security

OSV.dev

1 open1 fixed

CVE-2025-61260criticalCVSS 9.8

OpenAI Codex CLI enables code execution through malicious MCP (Model Context Protocol) configuration files

A vulnerability was identified in OpenAI Codex CLI v0.23.0 and before that enables code execution through malicious MCP (Model Context Protocol) configuration files. The attack is triggered when a user runs the codex command inside a malicious or compromised repository. Codex automatically loads project-local .env and .codex/config.toml files without requiring user confirmation, allowing attackers to embed arbitrary commands that execute immediately.

Affected: >= 0source →

CVE-2025-59532high · fixed

Codex has sandbox bypass due to bug in path configuration logic

Due to a bug in the sandbox configuration logic, Codex CLI could treat a model-generated `cwd` as the sandbox’s writable root, including paths outside of the folder where the user started their session. This logic bypassed the intended workspace boundary and enables arbitrary file writes and command execution where the Codex process has permissions - this did not impact the network-disabled sandbox restriction. **Remediation** We released a patch in Codex CLI **0.39.0** that canonicalizes and

Affected: >= 0.2.0Fixed in 0.39.0source →

Community

Reviews

Be the first to review

Have you used this server?

Share your experience — it helps other developers decide.

How easy was setup?Did it work reliably?How was the documentation?

Frequently Asked Questions

Is Agent Collab safe to use?: Agent Collab has 1 known CVE tracked by MCPpedia. You can verify these on OSV.dev by searching for "@openai/codex". Review the Security section above for details before installing.
How do I install Agent Collab?: Agent Collab supports copy-paste install configs on its MCPpedia page for Claude Desktop, Cursor, and Claude Code. Scroll to the Quick Install section and select your client.
What AI clients work with Agent Collab?: Agent Collab is compatible with claude-desktop, cursor, claude-code. It uses stdio and sse transport.
Is Agent Collab actively maintained?: Agent Collab is actively maintained — last commit was 8 days ago. It has 32 GitHub stars.

Similar servers

Others in ai-ml

View all →

Dsh MCP Client97

MCP client bridge: connects to MCP servers and registers their tools on ctx.tools

190.4k 4

Gpt Researcher97

An autonomous agent that conducts deep research on any data using any LLM providers

29.1k 5

io.github.mims-harvard/tooluniverse96

1000+ scientific tools for AI scientists: life science, reserach, literature, and more.

1.7k 4

Antigravity Workspace Template96

Workspace template + MCP server for Claude Code, Codex CLI, Cursor & Windsurf. Multi-agent knowledge engine (ag-refresh / ag-ask) that turns any codebase into a queryable AI assistant.

1.3k 2

MCP Security Weekly

Get CVE alerts and security updates for Agent Collab and similar servers.

Community

Discussion

Start a conversation

Ask a question, share a tip, or report an issue.

Has anyone used this with Cursor?How do you handle auth?Any alternatives?

Edit this pageView history

AI / ML

Step 1

Install in your client

Config is the same across clients — only the file and path differ.

Supported in Claude Desktopstdio, sse · Node 18+

Paste into ~/Library/Application Support/Claude/claude_desktop_config.json

{
  "mcpServers": {
    "agent-collab": {
      "args": [
        "-y",
        "@openai/codex"
      ],
      "command": "npx"
    }
  }
}

Are you the author?

Add this badge to your README to show your security score and help users find safe servers.

Embed in your READMEAbout badges →

[![MCPpedia Score](https://mcppedia.org/api/badge/agent-collab)](https://mcppedia.org/s/agent-collab)

Read me

What Agent Collab does

Delegate work from Claude Code to other AI coding agents.

Test This Server

Run this in your terminal to verify the server starts. Then let us know if it worked — your result helps other developers.

npx -y '@openai/codex' 2>&1 | head -1 && echo "✓ Server started successfully"

After testing, let us know if it worked:

README

agent-collab

Delegate work from Claude Code to other AI coding agents.

Run Codex and Cursor as background workers inside Claude Code — reviews, tasks, and rescues with tracked jobs, live progress, and session handoffs.

Quick start • Codex plugin • Cursor plugin • Ambient delegation • Goal plugin • How it works • Development • License

One marketplace, three plugins:

Plugin	Delegates to	Highlights
`codex`	OpenAI Codex (`codex` CLI)	`/codex:task`, `/codex:review`, `/codex:adversarial-review`, `/codex:rescue`, session transfer, optional stop-review gate. Community fork of openai/codex-plugin-cc with Windows fixes.
`cursor`	Cursor (`cursor-agent` CLI)	`/cursor:review`, `/cursor:task`, ownership-verified cancel, WSL support on Windows, resolved-model + token-usage recording.
`goal`	Long-horizon goals	`/goal:set`, `/goal:step` — one increment per invocation, delegating through the other two plugins, with mechanical state and dispositions.

The codex and cursor plugins share the same job chassis: background jobs with progress-aware status, stored result output, model recording, resume handoffs into the native tool, and cancel that never kills a process it can't prove it owns.

Quick start

Add the marketplace in Claude Code:

/plugin marketplace add zebbern/agent-collab

Install any of the plugins:

/plugin install codex@agent-collab
/plugin install cursor@agent-collab
/plugin install goal@agent-collab

Reload, then check readiness:

/reload-plugins
/codex:setup
/cursor:setup

A first run that starts durable work you can collect after restarting Claude:

/codex:task --background --fresh map this repository's architecture and identify the main runtime entrypoints

Copy the returned job ID, close Claude, then reopen the same repository and run:

/codex:status <job-id> --wait
/codex:result <job-id>

For the official OpenAI plugin instead of this fork, use /plugin marketplace add openai/codex-plugin-cc.

Codex plugin

[!IMPORTANT] This is an unofficial fork. It is not built, endorsed, or supported by OpenAI. It is a modified version of openai/codex-plugin-cc at v1.0.6, maintained by @zebbern. Report issues here, not to OpenAI.

What this fork changes (see the changelog for detail)

Windows correctness: the broker ownership registry works on win32 (POSIX permission checks gated off), registry lock contention handles Windows EPERM rename semantics, taskkill/tasklist bypass MSYS argument mangling, and the test suite passes on Windows (upstream fails 9 of its own tests there).
Durability: state.json is written atomically under an exclusive lock; corrupt state/job files are quarantined and rebuilt; terminal jobs can never be resurrected by racing writers.
Observability: /codex:status is progress-aware (file changes, command executions, token usage, likely dead detection); jobs record the model and reasoning effort they ran

... View full README on GitHub

Loading README…

Scored, not listed

Why this score

Five weighted categories — click any category to see the underlying evidence.

Score breakdown

57/100across 5 weighted dimensions

How we score →

0255075100

−43

Security

Maintenance

Efficiency

Documentation

Compatibility

Categoriesclick a row to see evidence

Security

OSV.dev

1 open1 fixed

CVE-2025-61260criticalCVSS 9.8

OpenAI Codex CLI enables code execution through malicious MCP (Model Context Protocol) configuration files

Affected: >= 0source →

CVE-2025-59532high · fixed

Codex has sandbox bypass due to bug in path configuration logic

Affected: >= 0.2.0Fixed in 0.39.0source →

Community

Reviews

Be the first to review

Have you used this server?

Share your experience — it helps other developers decide.

How easy was setup?Did it work reliably?How was the documentation?

Frequently Asked Questions

Is Agent Collab safe to use?: Agent Collab has 1 known CVE tracked by MCPpedia. You can verify these on OSV.dev by searching for "@openai/codex". Review the Security section above for details before installing.
How do I install Agent Collab?: Agent Collab supports copy-paste install configs on its MCPpedia page for Claude Desktop, Cursor, and Claude Code. Scroll to the Quick Install section and select your client.
What AI clients work with Agent Collab?: Agent Collab is compatible with claude-desktop, cursor, claude-code. It uses stdio and sse transport.
Is Agent Collab actively maintained?: Agent Collab is actively maintained — last commit was 8 days ago. It has 32 GitHub stars.

Similar servers

Others in ai-ml

View all →

Dsh MCP Client97

MCP client bridge: connects to MCP servers and registers their tools on ctx.tools

190.4k 4

Gpt Researcher97

An autonomous agent that conducts deep research on any data using any LLM providers

29.1k 5

io.github.mims-harvard/tooluniverse96

1000+ scientific tools for AI scientists: life science, reserach, literature, and more.

1.7k 4

Antigravity Workspace Template96

Workspace template + MCP server for Claude Code, Codex CLI, Cursor & Windsurf. Multi-agent knowledge engine (ag-refresh / ag-ask) that turns any codebase into a queryable AI assistant.

1.3k 2

MCP Security Weekly

Get CVE alerts and security updates for Agent Collab and similar servers.

Community

Discussion

Start a conversation

Ask a question, share a tip, or report an issue.

Has anyone used this with Cursor?How do you handle auth?Any alternatives?