Monitor Zed with FrontierStack
Zed is a fast, open-source code editor with AI built in: its Agent Panel runs Zed's own agent on Zed's hosted models, your own API keys or local models through Ollama, Zeta predicts your next edit, and external agents such as Claude Code, Codex and Gemini CLI run inside Zed over the Agent Client Protocol. Zed's agent reads MCP servers from context_servers in ~/.config/zed/settings.json. In the MCP pane, Add to Zed adds FrontierStack's credential-free bridge there with its own agent key. Zed's settings usually carry comments, so FrontierStack edits the file in place: your comments and formatting stay, the original is kept once as a backup, and a file it can't read safely is left alone. It also puts the FrontierStack skill in ~/.agents/skills, where Zed's agent finds personal skills. Zed asks before each FrontierStack tool unless you allow it. FrontierStack can also be an agent in Zed itself: turn on Answer in Zed's agent panel and Zed lists FrontierStack next to its own agent, Claude Code and Codex, so you can start a thread with it and ask about your servers while you code. Only the text you type is sent. It can only look until you also turn on Allow changes from Zed, and even then every change waits for approval in FrontierStack, never in Zed.
Run it from your Mac.
FrontierStack installs, monitors and secures services on this Mac and on linked servers; and keeps your passwords and keys out of any AI's sight.
Download FrontierStackApple notarized · Safe & secure · macOS 13 Ventura+
