CVE WatchJul 24, 2026

CVE-2026-33234

AutoGPT is a workflow automation platform for creating, deploying, and managing continuous artificial intelligence agents. In versions 0.1.0

CVSS

5.0

Medium

EPSS

0.3%

p23

KEV

—

Exploit Today

0-100

Published: May 19, 2026 · Last modified: Jul 24, 2026 · CWE-918

EPSS · 30d

0.3%EPSS · 30 days0.3%

2026-08-122026-09-09

Technical description

AutoGPT is a workflow automation platform for creating, deploying, and managing continuous artificial intelligence agents. In versions 0.1.0 through 0.6.51, SendEmailBlock in autogpt_platform/backend/backend/blocks/email_block.py accepts a user-supplied smtp_server (string) and smtp_port (integer) as per-execution block inputs, then passes them directly to Python's smtplib.SMTP() to open a raw TCP connection with no IP address validation. This completely bypasses the platform's hardened SSRF protections in backend/util/request.py — the validate_url_host() function and BLOCKED_IP_NETWORKS blocklist that every other block uses to block connections to private, loopback, link-local, and cloud metadata addresses. An authenticated user on a shared AutoGPT deployment can use this to perform non-blind internal network port scanning and service fingerprinting: smtplib reads the target's TCP banner on connect and embeds it in the exception message, which is persisted as user-visible block output via the execution framework. This issue has been fixed in version 0.6.52.

Official references

Related CVEs

CVECVSSEPSSKEVRExploitTitleMod.

CVE-2026-19233—

—

——0CWE-918: Server-Side Request Forgery (SSRF) vulnerability exists that could cause unauthorized command execution and disclosure of server data when an attacker with a privileged account sends crafted, unvalidated parameters to a server endpoint.1d CVE-2026-867717.6 HIG

—

——0Snipe-IT versions before 8.7.0 fail to HTML-escape the employee_num field in the acceptance PDF generator, allowing attackers with users.edit permission to inject img tags into TCPDF's writeHTML() function. Attackers can craft a malicious employee_num value containing an img tag with an arbitrary HTTP(S) URL to trigger server-side requests to internal services, cloud metadata endpoints, or external targets when a victim signs an asset acceptance.1d CVE-2026-878217.1 HIG

—

——0Lara Dashboard through 1.3.1 contains a server-side request forgery vulnerability in the POST /api/admin/builder/markdown/fetch endpoint that allows any authenticated user to fetch arbitrary URLs and read the response body. Attackers can read internal HTTP services and cloud metadata including IAM credentials by supplying malicious URLs without host validation or redirect restrictions.1d CVE-2026-796357.3 HIG

—

——0Dell SCG 5.0 Appliance versions prior to 5.36.00.16 and Dell SCG 5.0 Application versions prior to 5.36.00.00, contains a Server-Side Request Forgery (SSRF) vulnerability. An unauthenticated attacker with remote access could potentially exploit this vulnerability, leading to unauthorized access.1d CVE-2026-57866—

—

——0Server side request forgery in Apache Impala versions 4.4.x and 4.5.x. Authenticated Impala users with permissions to execute the ai_generate_text() function can exfiltrate secrets provided by the credential providers configured in the `hadoop.security.credential.provider.path` property of `core-site.xml`. The secret's key must be known to the user.1d CVE-2026-54048—

—

——0Specifying tblproperties('avro.schema.url'=' http://...' ) or with a 'file:///' URI on a table in Impala 2.0.0 to 4.5.1 on all platforms allows an attacker to trigger a GET request to internal endpoints they may not have access to but that Impala does and the response my be exposed via parsing error messages. Users are recommended to upgrade to version 4.5.2, which fixes this issue.1d