Postman silently auto-installs AI plugins and attempts to share all API data

Your AI coding assistant may now route all your API work through Postman without your explicit permission.

Postman's desktop application has been observed to automatically install AI plugins and integrate with other coding agents without explicit user consent. Upon opening Postman version 12.33.4, users reported unexpected modifications to their coding environments, including the installation of a Postman CLI, a Claude Code plugin, and a Codex Postman plugin. Log files indicate an "auto-existing-upgrade" process that successfully installed these components. The Postman plugin also altered the global configuration of the Codex agent, enabling a Postman marketplace and a startup hook that directs broad API engineering tasks through Postman skills. Furthermore, the installer code attempts to automatically approve its own hook, a process that typically requires explicit user trust. A remote MCP connection was also enabled, though it failed to authenticate. These changes occurred on October 24, 2026, affecting users on macOS and impacting both Codex and Claude Code integrations, with a Postman executable also found in the user's local bin directory.

The silent, automatic installation of AI plugins and configuration changes by Postman raises significant concerns about user consent and security boundaries in software integration. By default, opening Postman now appears to trigger a broad directive for AI agents to route API engineering tasks through Postman skills, overriding user control over their development workflows. This bypasses the expected user opt-in for such deep system integrations, suggesting a potential shift in how application ecosystems are designed to interact with user data and AI tooling.

This incident highlights a growing tension between seamless integration and user privacy, particularly as AI tools become more deeply embedded in developer workflows. The lack of explicit consent mechanisms for installing plugins, configuring hooks, and potentially sharing API data, could erode trust across the developer ecosystem. Future developments will likely focus on stricter regulations and user-demand for transparency and granular control over AI integrations and data sharing within development tools.