Critical Flaws Patched in HashiCorp Terraform MCP, Veeam Console, and Django
Three major software providers – **HashiCorp**, **Veeam**, and the **Django Software Foundation** – have released urgent patches for a total of 11 vulnerabilities. These include critical issues that could lead to unauthenticated credential theft, cross-tenant data exposure, and remote code execution, impacting IT security professionals and privacy-conscious users managing these platforms.
A recent wave of security advisories has highlighted critical vulnerabilities across widely used enterprise and web development platforms. **HashiCorp Terraform MCP Server**, **Veeam Service Provider Console**, and **Django** have all received patches for serious flaws, some with CVSS scores reaching 10.0.
Operators are strongly advised to update their systems immediately to the recommended versions to mitigate potential risks.
## Veeam Console: Unauthenticated Credential Theft and RCE Risks
**Veeam Service Provider Console (VSPC)**, a multi-tenant solution for managing customer backups, received four fixes in build 9.3.0.35057. Two of these are critical:
* **CVE-2026-58073** (CVSS score: 9.5) allows an unauthenticated attacker to impersonate a managed agent and obtain its credentials. While its attack complexity is rated high, the lack of authentication makes it a significant concern.
* **CVE-2026-58072** (CVSS score: 9.0) is an arbitrary file write vulnerability on the management server, which could lead to remote code execution. This flaw requires a low-privilege account.
Two additional high-severity bugs, **CVE-2026-58067** (unauthenticated memory-exhaustion DoS) and **CVE-2026-58071** (proxied appliance API exposure), complete the set. All four vulnerabilities affect VSPC 9.2.1.33875 and all earlier version 9 builds. The fix is an upgrade to 9.3.0.35057.
This marks the second critical patch cycle for the console in approximately three months, following the resolution of **CVE-2026-32998** (a 9.4-rated RCE bug) in May.
## HashiCorp Terraform MCP Server: Cross-Tenant Credential Reuse
**HashiCorp's Terraform MCP Server**, which facilitates AI assistant integration with Terraform via the Model Context Protocol, addressed three related flaws in its Streamable HTTP transport. These were disclosed on July 28 and fixed in version 1.1.0 (released July 14).
Deployments running solely in stdio mode (local, single-user setups) are unaffected. The vulnerabilities reside in the multi-user HTTP mode, designed for centralized, shared deployments.
* **CVE-2026-16498** (CVSS score: 10.0) is a critical cross-tenant credential-reuse bug affecting stateless HTTP mode. The underlying MCP library failed to assign unique session identifiers, allowing one user's Terraform token to be reused for subsequent users' requests.
* **CVE-2026-16496** (CVSS score: 8.9) is a similar isolation failure in stateful mode, the default for central deployments. This flaw allowed a user who obtained another user's session ID to execute tool calls with that user's Terraform client, accessing resources permitted by the victim's token. This vulnerability was reported by **Juan Pablo Martinez Kuhn** of **Coinspect**.
* **CVE-2026-14869** (CVSS score: 8.6) is a server-side request forgery (SSRF) flaw. An unauthenticated caller able to reach the Streamable HTTP listener could make the server send its configured bearer token to an attacker-controlled endpoint.
**HashiCorp** recommends upgrading to version 1.1.0 or later. For those unable to upgrade immediately, restricting network access to the Streamable HTTP listener to trusted users and treating MCP session IDs as sensitive values is crucial.
## Django: GeoDjango RCE and Other Fixes
**Django** released versions 6.0.8 and 5.2.17 on August 4, patching four CVEs. The most severe, **CVE-2026-15307**, is rated high and affects **GeoDjango**, the framework's geographic-data layer.
* **CVE-2026-15307** could allow a staff user with view permission on a registered model containing a spatial field to write a file to disk or make the Django process issue a network request via spatial lookups. Depending on the raster driver and subsequent application imports, this could lead to remote code execution.
Other patched flaws include:
* **CVE-2026-15920**: A moderate stored cross-site scripting (XSS) bug in the admin interface.
* **CVE-2026-15830**: A moderate denial of service through deeply nested GEOMETRYCOLLECTION objects.
* **CVE-2026-15337**: A low-severity memory-consumption denial of service in `check_for_language()`.
Older, unsupported Django branches (5.1, 5.0, and 4.2) were not evaluated but may also be affected. **Django's** GIS code has previously attracted attacker attention this year, with **CVE-2026-1207** (a SQL injection flaw in PostGIS raster lookups) being exploited in the wild, as reported by **CrowdSec**.
## Mitigation and Next Steps
None of the advisories indicate active exploitation for these newly disclosed vulnerabilities, and none appear in **CISA's Known Exploited Vulnerabilities catalog** as of August 5, 2026. However, the severity of these flaws necessitates immediate action.
IT security professionals and system administrators should prioritize patching their **Terraform MCP Server** to 1.1.0+, **Veeam Service Provider Console** to 9.3.0.35057, and **Django** to 6.0.8 or 5.2.17 to protect against potential attacks. Regular security audits and staying informed about vendor advisories remain critical practices.