CVE-2026-111079

Goose terminal history can disclose secrets to model providers

Goose project / Red Hat CNA · Goose terminal integration

First documented disclosure · Review checkpoint

Summary

Terminal integration sends recent shell commands to the configured LLM endpoint without redaction.

Severity and exploitation evidence

High · 7.4/10 · CVSS 3.1. Severity source.

Exploitation not established: The reviewed sources do not establish malicious exploitation. A published proof of concept is not evidence of a customer compromise.

Technical severity is not a measure of observed attacks or of risk to your deployment. Attempts, demonstrations, and successful compromise retain the qualifications in the source.

How the weakness works

Secret-bearing history crosses from a local shell into provider requests and logs.

Conditions required

Terminal integration enabled and sensitive values present in recent commands.

Potential impact

Potential exposure of credentials and secret-bearing environment assignments.

Affected versions and configurations

Goose terminal-history integration; the CNA does not provide a complete upstream version range.

Fix and mitigation status

1.56.0 changes new history hooks to opt-in; it does not redact opted-in history.

Practical next steps

Upgrade, start a fresh shell with history sharing disabled, inspect existing hooks, and rotate credentials that were transmitted.

Operating lesson

Minimize model context and keep credentials out of shell command history.

Evidence and limits

The October 21 public hardening PR predates the October 22 CVE publication. Opt-in is mitigation, not proof that earlier disclosures were erased.

Related incident and research records

Related records can include laboratory demonstrations; a relationship does not itself establish exploitation.

Sources