Postman installs plugins into Claude Code and Codex without explicit consent, developers say

Users report that the Postman desktop app adds plugins to Claude Code and Codex on its own. Postman says it is a convenience bundle and promised a clearer opt-in.

A GitHub issue that has passed 400 points on Hacker News describes what happened on one Mac after a Postman update. The desktop app installed the Postman CLI, enabled a plugin in Claude Code and added an entry to the Codex configuration, including a remote MCP server. The reporter says this hit already signed-in users and nobody asked them first.

The startup hook drew the most attention. According to the report, it injects an instruction into each agent session telling it to load Postman's skills for non-trivial API work. The reporter also argues the installer writes a trust record for Postman's own hook, which they say is their reading of the installer source. Other users confirm similar behaviour on different versions and mention Cursor and Copilot.

What the report does not claim matters too. The author states it does not establish data exfiltration, and other participants say they have no evidence anything was uploaded. In one case the remote MCP server only answered that authentication was required. The risk is therefore mainly that a tool changes the environment of your AI agents without anyone approving it.

A Postman representative wrote on 3 November that the CLI and agent plugins are bundled for users who struggled to install them separately, that this is mentioned on the download page and in the update flow, and that components can be uninstalled in the app. They promised a clearer opt-in for existing users and a review of the hook behaviour. Commenters call that insufficient and ask for opt-in for everyone.