AIToday
Large Language ModelsAI Coding AssistantsAI WeeklyPublished: Nov 4, 2026, 16:00 JST

Postman silently installs plugins into Codex and Claude Code

Send on LINE
Postman silently installs plugins into Codex and Claude Code

3 Key Points

  1. What happened

    Opening Postman desktop on October 24 installed a plugin into Codex and Claude Code, and the installer then wrote the trust record for its own startup hook; GitHub issue #14370 was filed October 25.

  2. Why it matters

    The installer granted its own startup hook persistent trust inside the coding agent, sidestepping the very approval prompt meant to keep outside parties from running code — a hedge: this appears to be intentional design rather than a shipping accident.

WHO IT HITSDevelopers running Postman desktop alongside coding agents like Codex or Claude Code are directly affected, since opening the app rewrote global agent config files without their action; security teams reviewing agent permissions may also need to reassess what third-party apps can modify.

Not sure about something? Ask the AI

Questions and answers are published on this page.

Summaries like this, every morning.

Context & Analysis

The trigger appears to be an automatic upgrade path rather than a deliberate user action: the GitHub reporter's logs show the trigger as auto-existing-upgrade, with results claude-code:installed and codex:installed on Postman desktop 12.33.4. That detail matters because it suggests the plugin installation happened as a side effect of simply opening the app, not through any menu a developer chose.

The startup instruction embedded in the hook tells the coding agent to load the postman:api-engineer skill before responding to any non-trivial API engineering task — wording that routes API-shaped work through Postman even when the user never names Postman. Combined with the installer writing the hook's own trust record, the incident raises a question about where the boundary sits between an app configuring itself and an app configuring the tools around it.

The GitHub thread, which carries 99 thumbs-up reactions, asks Postman to halt automatic agent-plugin installation pending explicit opt-in, remove installer-side hook self-approval, and narrow the startup instruction to explicit user requests. Community reaction logged alongside the coverage was reported as 90% skeptical, with some healthcare firms decertifying Postman and developers migrating to Bruno, Insomnia and Yaak as alternatives, though Postman itself has not publicly responded.

FAQ
Did any of my data actually get sent to Postman?
According to the incident logs, the MCP remote endpoint at mcp.postman.com failed authentication, so no API collection data was transmitted despite the session being established.
Which files did the installer modify?
On macOS it touched ~/.codex/config.toml, ~/.claude/settings.json and ~/.claude/plugins/installed_plugins.json, and placed a Postman executable in ~/.local/bin.
Has Postman responded or fixed it?
Postman has not publicly responded, and the behavior has not yet been reproduced on a clean system by an independent party.
Send on LINE

AI news that matters for your work, delivered every morning.

Pick your industry and the AI tools you use, and get news related to your work every day.

Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →

Ask AI

Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.

Questions and answers are published on this page.

Related Articles

Next articleNissay Plus starts AI underwriting on Oct 30, 2026