No red flags found in any of the 11 categories — no credential harvesting, no data exfiltration, no curl-pipe-shell installer. Scanned against the SlowMist agent-security taxonomy, refreshed every 8 hours. Full audit →
by Riccardo8888 · MCP Server · ★ 55
Last updated: · Indexed by AgentSkillsHub · Auto-synced every 8h
🔒 Is agent-link safe to install? View the security audit →
agent-link Two coding agents, on two different machines, talking to each other directly. You and a colleague are both working with Claude Code, or one of you is on Codex. Instead of copying "I'll take the auth module, you take the frontend" back and forth by hand, the agents say it to each other. They split work, hand off tasks and report progress, in an end-to-end encrypted room. There is nothing to deploy and nothing to pay for. One private repo you can both push to is the entire infrastructure. No server, no account, no open ports. The repo can be the one you are already working in together. Nothing has to be created for this. One Python dependency, Python 3.10+, and . Windows, Linux and macOS. Install The repo is private for now: you need to be a collaborator and logged in to git (, an ssh key, or a PAT) before the install line will work. On each machine. Nothing to clone and nothing to keep: is not usually installed already: then , and reopen the shell. Plain works too, and then also works. Or from a clone, if you have one: powershell powershell -ExecutionPolicy Bypass -File install.ps1 # Windows pwsh -File install.ps1
| Stars | 55 |
| Forks | 7 |
| Language | Python |
| Category | MCP Server |
| License | MIT |
| Quality Score | 67.9803050771563/100 |
| Open Issues | 2 |
| Last Updated | 2026-08-21 |
| Created | 2026-08-11 |
| Platforms | claude-code, codex, mcp, python |
| Est. Tokens | ~17k |
These tools work well together with agent-link for enhanced workflows:
Looking for a agent-link alternative? If you're comparing agent-link with other mcp server tools, these 6 projects are the closest alternatives on Agent Skills Hub — ranked by topic overlap, star count, and community traction.
Close the lid and walk away — every window, desktop and terminal on your machines, live in any browser. E2E-en
DeepContext is an MCP server that adds symbol-aware semantic search to Claude Code, Codex CLI, and other agent
One workspace all your AI agents can read and write. Encrypted client-side, model-agnostic, gone 24h after the
Better 'AskUserQuestion' - A lightweight MCP server/OpenCode plugin/Agent Skills + CLI tool that allows your L
AI-to-AI communication protocol for Claude, Gemini, and other AI assistants
Portable SKILLs workspace for Claude Code, Codex, and Gemini — issue-driven workflows and cross-agent collabor
Explore other popular mcp server tools:
agent-link is An end-to-end encrypted channel between coding agents, anywhere. It is categorized as a MCP Server with 55 GitHub stars.
agent-link is primarily written in Python. It covers topics such as agent-communication, ai-agents, claude-code.
You can find installation instructions and usage details in the agent-link GitHub repository at github.com/Riccardo8888/agent-link. The project has 55 stars and 7 forks, indicating an active community.
agent-link is released under the MIT license, making it free to use and modify according to the license terms.
The top alternatives to agent-link on Agent Skills Hub include Reminal, deepcontext-mcp, vnsh. Each offers a different approach to the same problem space — compare them side-by-side by stars, quality score, and community activity.
Grades come from a rule-based scan built on the SlowMist agent-security taxonomy, covering 11 red-flag categories including credential harvesting, data exfiltration, and curl | sh installers. It is a first-layer scan, not a manual audit — we say so rather than overstate it.
The scale of the problem is documented independently: Liu et al. (2026), in a study of 31,132 agent skills, report that 26.1% contain security vulnerabilities. Our own full-catalog census is published as a citable open dataset.
Sources & who's responsible: