Skip to content

op_run tool leaks OP_SERVICE_ACCOUNT_TOKEN master credential to subprocesses and model context

High
CakeRepository published GHSA-q3gg-rgxh-gh64 Sep 1, 2026

Package

npm @takescake/1password-mcp (npm)

Affected versions

< 4.0.2

Patched versions

4.0.2

Description

Impact

In @takescake/1password-mcp versions prior to 4.0.2, the op_run tool spawned child processes by cloning the full server environment ({ ...process.env }). In standard configurations where OP_SERVICE_ACCOUNT_TOKEN (or macOS keychain credentials) is set in the server's environment, this exposed the master service account token to every command and subprocess executed by op_run.

Furthermore, output redaction was only applied to secret values resolved from explicit op:// references passed in the tool's env parameter. Consequently, any command or script that echoed its environment (such as env, printenv, or debug logging) returned the master service account token in plaintext to the model's conversation transcript.

Patched Versions

This vulnerability has been patched in @takescake/[email protected].

  • op_run now explicitly strips OP_SERVICE_ACCOUNT_TOKEN, OP_KEYCHAIN_SERVICE, and OP_KEYCHAIN_ACCOUNT from the child process environment before execution.
  • The server's master service account token is automatically included in the defense-in-depth redaction targets across stdout, stderr, spawn errors, and caught exceptions («REDACTED:OP_SERVICE_ACCOUNT_TOKEN»).
  • Full stream redaction is performed prior to the 5 MiB output truncation limit to prevent partial secret leakage across boundaries.

Workarounds

Users unable to upgrade to 4.0.2 immediately should avoid running commands via op_run that execute untrusted scripts or print environment variables.

References

Credit

Discovered and reported by independent security researcher Syed Anas Mohiuddin (@SyedAnas01).

Severity

High

CVE ID

No known CVE

Weaknesses

Exposure of Sensitive Information to an Unauthorized Actor

The product exposes sensitive information to an actor that is not explicitly authorized to have access to that information. Learn more on MITRE.

Improper Removal of Sensitive Information Before Storage or Transfer

The product stores, transfers, or shares a resource that contains sensitive information, but it does not properly remove that information before the product makes the resource available to unauthorized actors. Learn more on MITRE.

Credits