CVE-2026-113663

validatedZero produced an exploit or a check.

GHSA-2h4w-29cv-2cmv ↗EUVD-2026-117877 ↗

check availableZero validated a working check for it.not proven exploitableZero has a detection check, but hasn't produced a full working exploit yet.
cve.orgpublic pocExploit code has been published by someone. Source signals can also carry a PoC (Proof of Concept).kevAttacks confirmed in the real world.
ag2ai ag2 >= 0.13.2, < 1.1.1; >= 0.13.2, <= 0.14.1
fixed in 1.1.1
CWE-78 Improper Neutralization of Special Elements used in an OS Command ('OS Command Injection')
ETK-2026-3680Zero's internal tracking ID, independent of the CVE number., permanent identifier
pkg:pypi/ag2
pkg:pypi/autogen
+51mTTV
7mAI analysisAnalysis: how long the AI spent reversing this signal.
Check if you are vulnerable
timeline
0 = trigger signal
trigger signal (cve.org) +0mcheck available+51magent stopped
0
+10min
+20min
+30min
+40min

description

AG2 SandboxShellTool before version 1.1.1 and affected autogen versions can incorrectly validate shell commands when allowlist or read-only controls are enabled. An attacker able to influence an agent command may execute unintended commands. LocalEnvironment deployments can run those commands with the privileges assigned to the agent process.

mitigation

Upgrade ag2 to version 1.1.1 or later. Until upgraded, disable SandboxShellTool for attacker-influenced input and run agents in isolated, least-privileged environments.

fixed in 1.1.1

check
available
exploit
no
automationDeterministic: a check that runs on its own, the same request on every target. Contextual: an agentic technique that needs to be applied per target, for bugs that depend on contextual flows.
contextual
auth requiredWhether an attacker needs an account first, based on the CVSS vector's Privileges Required.
yes, low privileges
EPSSEstimate of the chance this CVE is exploited in the next 30 days. Refreshed daily.
–
agent run time
7m
cve.org record, t = 0The moment cve.org publishes the CVE in full.
2026-11-05 22:53Z
Δ to public pocExploit code has been published by someone. Source signals can also carry a PoC (Proof of Concept).
–
Δ to kevAttacks confirmed in the real world.
–

outcome

Contextual check: a technique exists and was validated, but no deterministic check could cover this one.

source signals

Public signals that triggered or updated the pipeline. These are inputs for the deltas above.

  1. 2026-11-05 22:53Z
    cve_org: trigger signal, started the run

this entry in the feed

GET /api/v1/feed/CVE-2026-113663

{
  "etk_id": "ETK-2026-3680",
  "cve_id": "CVE-2026-113663",
  "display_id": "CVE-2026-113663",
  "ghsa_id": "GHSA-2h4w-29cv-2cmv",
  "euvd_id": "EUVD-2026-117877",
  "status": "published",
  "anchor_state": "anchored",
  "title": "Shell command injection in AG2 SandboxShellTool",
  "summary": "AG2 SandboxShellTool before version 1.1.1 and affected autogen versions can incorrectly validate shell commands when allowlist or read-only controls are enabled. An attacker able to influence an agent command may execute unintended commands. LocalEnvironment deployments can run those commands with the privileges assigned to the agent process.",
  "mitigation": "Upgrade ag2 to version 1.1.1 or later. Until upgraded, disable SandboxShellTool for attacker-influenced input and run agents in isolated, least-privileged environments.",
  "vendor": "ag2ai",
  "product": "ag2",
  "affected": [
    {
      "affected_notes": [],
      "affected_ranges": [
        ">= 0.13.2, < 1.1.1"
      ],
      "ecosystem": "pip",
      "fixed_versions": [
        "1.1.1"
      ],
      "package_name": "ag2",
      "purl": "pkg:pypi/ag2"
    },
    {
      "affected_notes": [],
      "affected_ranges": [
        ">= 0.13.2, <= 0.14.1"
      ],
      "ecosystem": "pip",
      "fixed_versions": [],
      "package_name": "autogen",
      "purl": "pkg:pypi/autogen"
    }
  ],
  "cwes": [
    78
  ],
  "cvss_score": 8.8,
  "cvss_src": "cna_vendor",
  "cvss_vector": "CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H",
  "epss_score": null,
  "epss_percentile": null,
  "epss_asof": null,
  "automation": "contextual",
  "check": true,
  "check_confidence_basis": "assessed",
  "exploit": false,
  "pipeline_minutes": 7.046989083333333,
  "t0_at": "2026-11-05T22:53:40.444000+00:00",
  "signals": {
    "check": {
      "at": "2026-11-05T23:45:05.933726+00:00",
      "days": 0.035711686643518516,
      "is_lower_bound": false,
      "measured_at": "2026-11-05T22:53:40.444000+00:00",
      "url": null
    },
    "trigger_signal": {
      "at": "2026-11-05T22:53:40.444000+00:00",
      "days": 0,
      "is_lower_bound": false,
      "measured_at": "2026-11-05T22:53:40.444000+00:00",
      "url": "https://www.cve.org/CVERecord?id=CVE-2026-113663"
    }
  }
}

published 2026-11-05 23:37Z
last change 2026-11-06 04:20Z

Need the technical details behind an entry?

Every entry here shows what we attempted and when. The technical details, the generated exploit that fired it and the evidence behind the verdict are not public. Tell us about your use case and we’ll get back to you about extended API access.