Zero Radar
← feed
CriticalVulnerabilityCVE-2026-109444updated 25 Oct 2026, 15:20 UTC

OpenSRE can run shell commands without approval, exposing credentials

OpenSRE, an AI tool that helps investigate system alerts, can be tricked into running commands on the server without asking for permission. An attacker could hide instructions inside an alert or Slack message to steal passwords and keys.

Official, 2 sources
SeverityCriticalCVSS 9.3 of 10
Being exploited?UnclearNo reports either way yet
Fix available?UnclearNo fixed version published in the sources yet
How sure is this?Official, 2 sourcesBacked by an official source

How the attack works simplified

  1. 1Attacker plants instructionsHidden text placed in an alert or Slack message
  2. 2Agent triages messageOpenSRE reads the untrusted content
  3. 3Approval gate skippedexecution_allowed() always returns True
  4. 4Shell command runsshell_run executes on the host
  5. 5Credentials exfiltratedLLM, Slack, and cloud secrets are sent out
  6. ResultAttacker steals credentials and controls host

What happened

In OpenSRE through 0.1.2026.10.19, the headless presenter's execution_allowed() function always returns True, so the approval gate for shell_run commands is bypassed. Attacker-supplied text in triaged alerts or Slack messages can cause arbitrary shell execution on the host.

Why it is dangerous: A successful attack can run commands on the host and exfiltrate LLM, Slack, and cloud credentials, potentially giving attackers access to connected cloud accounts and messaging systems.

Are you affected?

You are affected if you run OpenSRE through version 0.1.2026.10.19 in headless mode and it processes alerts or Slack messages that an attacker could influence.

OpenSRE through 0.1.2026.10.19 (headless presenter)

What to do

  1. Check whether you run OpenSRE in headless mode on version 0.1.2026.10.19 or earlier
  2. Restrict or disable shell_run until a fixed version is available
  3. Rotate LLM, Slack, and cloud credentials that OpenSRE can access if exposure is suspected
  4. Watch the official advisory (GHSA-fqh3-76w5-5vvf) for a patched release

Timeline

  1. 20 Oct 2026, 01:17 UTC
    Publicly disclosed
  2. 20 Oct 2026, 01:17 UTC
    Reported by NVD · NVD
  3. 25 Oct 2026, 15:20 UTC
    GitHub Advisory reported on it · GitHub Advisory

Sources 2 reports, oldest first

Official vendor, government CERT or advisory database

Technical details

ID
CVE-2026-109444 · GHSA-fqh3-76w5-5vvf
CVSS
9.3 / 10
EPSS
0.7% chance of exploitation within 30 days
Disclosed
20 Oct 2026, 01:17 UTC
On Zero Radar
before the radar started
Last change
25 Oct 2026, 15:20 UTC

The plain-language parts and the attack diagram were written by an AI model (anthropic/claude-haiku-5.5) from the sources above. Facts like versions, exploitation and dates come from the sources themselves. How Zero Radar works

OpenSRE can run shell commands without approval, exposing credentials · Zero Radar