Dev builds MCP layer so Claude Code can work on servers without ever holding SSH keys
NETMESAFE · reddit · 2026-10-16
A security engineer built SSHive, an SSH client with an MCP layer that lets Claude Code work on servers without ever holding keys or its own shell. The model operates only inside sessions the human opened: it can run commands, read logs, and edit files; opening new connections is a separate permission, specific servers (like prod) can be made unreachable, and it can only close sessions it created. The MCP server runs on localhost with the bearer token in the macOS Keychain — no cloud relay. Typical flow: ask why nginx returns 502 on staging, it inspects the service, reads error logs, and proposes a fix while the human decides. macOS only for now.
More from coding & agent
- Reflection AI ships Beam 1.0 as engineer touts small-team velocity — Scobleizer · 2026-10-17
- Open-source MCP Lets Claude Generate Image & Video at Provider Price — Demo Cost $0.10 — Master_Relative5687 · 2026-10-17
- Measuring agents by wall-clock time: harness and multi-agent overhead matter — rickasaurus · 2026-10-17
- Spreadsheet Agent Allowlists Are Not Acceptance Tests — Define What Must Stay Unchanged — Competitive-Mud7832 · 2026-10-17
- Researchers find rogue Tencent agent fleet probing Amap, a new AI security incident — luisdans · 2026-10-17
- Thorsten Ball questions pull requests in the AI coding era: "Why are you still using them?" — kieranklaassen · 2026-10-17