38 vulnerabilities across 14 products scored HIGH or above on August 04, 2026.

  • 🟠 HIGH: 38

Exploit Status Upgrades

The following CVEs from previous bulletins have been upgraded based on new exploit intelligence:

  • [UPGRADED] CVE-2026-67305 (freerdp/freerdp) — F1: exploitable → functional, AAS: 10.0 → 12.0 (HIGH → CRITICAL). Originally in 2026-08-01 bulletin.
  • [UPGRADED] CVE-2026-15988 (suspended_starter/ai_engine) — F1: exploitable → functional, AAS: 9.6 → 11.6 (HIGH → HIGH). Originally in 2026-08-01 bulletin.
  • [UPGRADED] CVE-2026-67320 (axios/axios) — F1: theoretical → poc, AAS: 9.4 → 11.9 (HIGH → HIGH). Originally in 2026-08-01 bulletin.
  • [UPGRADED] CVE-2026-48449 (adobe/campaign_classic) — F1: exploitable → functional, AAS: 10.2 → 12.8 (HIGH → CRITICAL). Originally in 2026-07-30 bulletin.

🟠 [HIGH] cinnamon/kotaemon

1 CVE | CVSS 4.0: 9.3 | AAS 11.6

  • cpe:2.3:a:cinnamon:kotaemon:*:*:*:*:*:*:*:*

Cinnamon kotaemon through version 0.12.0 is affected by one critical vulnerability (CVE-2026-69098, CVSS 9.3 HIGH) involving insecure deserialization in the check_connection endpoint. An unauthenticated attacker can supply crafted input with a manipulated type field to instantiate arbitrary Python classes, enabling remote code execution with application process privileges without any authentication required.

Teams running kotaemon for document retrieval or RAG workflows should treat this as urgent. Review the vendor advisory at the linked GitHub issue for remediation guidance, and restrict network access to affected instances until a patched release is available.

Vendor Advisory

🟠 [HIGH] opensips/opensips

4 CVEs | CVSS 3.1: 9.8 | AAS 11.4

  • cpe:2.3:a:opensips:opensips:*:*:*:*:*:*:*:* (>= 4.0.0-alpha, < 4.0.0-rc1)
  • cpe:2.3:a:opensips:opensips:*:*:*:*:*:*:*:* (>= 3.4.0-beta, < 3.6.6)
  • cpe:2.3:a:opensips:opensips:*:*:*:*:*:*:*:* (>= 4.0.0-beta, < 4.0.1)
  • cpe:2.3:a:opensips:opensips:*:*:*:*:*:*:*:* (>= 3.4.0, < 3.6.6)

OpenSIPS version 4.0.0 and prior is affected by four vulnerabilities, including at least one rated CVSS 9.8 HIGH (CVE-2026-45538), involving a stack buffer overflow triggered when processing SIP messages with oversized header names via the sip_to_json() function. These flaws can be exploited by remote unauthenticated attackers sending crafted SIP messages, potentially leading to code execution or denial of service on exposed SIP infrastructure.

Organizations running OpenSIPS for VoIP or unified communications should prioritize remediation immediately. Review the vendor security advisories linked from the GitHub advisory for patches or updated releases, and consider restricting SIP endpoint exposure at the network level until fixes are applied.

Vendor Advisory

🟠 [HIGH] flowiseai/flowise

12 CVEs | CVSS 4.0: 9.5 | AAS 11.3

  • cpe:2.3:a:flowiseai:flowise:*:*:*:*:*:*:*:* (< 3.1.3)

FlowiseAI Flowise prior to version 3.1.3 is affected by twelve vulnerabilities, including multiple rated CVSS 9.5 HIGH, spanning code injection, input validation bypasses, and unsafe data interpolation across several agent components. The lead issue involves the CSVAgent directly embedding attacker-controlled data into executable Python code without adequate sanitization, enabling authenticated attackers to achieve arbitrary code execution on the server.

Teams deploying Flowise for LLM workflow orchestration should upgrade to version 3.1.3 or later immediately. Review the referenced pull request and associated advisories for full details, and audit any internet-exposed Flowise instances for signs of exploitation.

Vendor Advisory

🟠 [HIGH] veeam/one

2 CVEs | CVSS 4.0: 10.0 | AAS 11.3

  • cpe:2.3:a:veeam:one:*:*:*:*:*:*:*:*

Veeam ONE is affected by two vulnerabilities, including at least one rated CVSS 10.0 HIGH, enabling remote unauthenticated code execution on the agent host. Veeam products are heavily targeted by ransomware operators, making these flaws especially dangerous for organizations relying on Veeam for backup and monitoring infrastructure.

Administrators should apply patches immediately by following the guidance in Veeam KB4892. Given the unauthenticated remote code execution vector and the critical role Veeam ONE plays in enterprise backup environments, treat this as an emergency remediation priority.

Vendor Advisory

🟠 [HIGH] qualcomm,_inc./snapdragon

1 CVE | CVSS 3.1: 9.6 | AAS 10.7

  • cpe:2.3:a:qualcomm:snapdragon:*:*:*:*:*:*:*:*

Qualcomm Snapdragon chipsets are affected by one vulnerability (CVE-2026-25289, CVSS 9.6 HIGH) involving memory corruption when processing malformed NAN (Neighbor Awareness Networking) Service Discovery Frames with invalid length values in Device Capability Extended attributes. This is a wireless-proximity attack vector that requires no authentication, potentially affecting a wide range of mobile devices, IoT equipment, and embedded systems using Snapdragon platforms.

Device manufacturers and enterprise mobility teams should review the Qualcomm August 2026 Security Bulletin and apply firmware updates as they become available from OEMs. Organizations with large mobile fleets or Qualcomm-based IoT deployments should prioritize patch scheduling given the severity and over-the-air exploitability of this flaw.

Vendor Advisory

🟠 [HIGH] google/android

1 CVE | CVSS 3.1: 9.8 | AAS 10.6

  • cpe:2.3:a:google:android:*:*:*:*:*:*:*:*

Google Android is affected by one vulnerability (CVE-2026-0163, CVSS 9.8 HIGH) involving a use-after-free condition in the VPU kernel driver (vpu_ioctl.c) that enables remote privilege escalation without user interaction or additional execution privileges. This flaw is particularly severe as it requires no user interaction and affects Pixel devices at the kernel driver level.

Organizations managing Android device fleets, especially those with Pixel devices, should apply the August 2026 Pixel Security Bulletin update immediately. Given the remote exploitability and lack of user interaction required, this should be treated as a high-priority patch cycle for enterprise mobility and MDM teams.

Vendor Advisory

🟠 [HIGH] nvidia/dynamo

3 CVEs | CVSS 3.1: 9.8 | AAS 10.6

  • cpe:2.3:a:nvidia:dynamo:*:*:*:*:*:*:*:*

NVIDIA Dynamo for Linux is affected by three vulnerabilities, including at least one rated CVSS 9.8 HIGH (CVE-2026-24254), involving an out-of-bounds write in the multimodal serving topology that can lead to code execution, privilege escalation, data tampering, denial of service, and information disclosure. These flaws are particularly concerning for organizations using Dynamo to serve AI and large language model inference workloads in production environments.

Teams deploying NVIDIA Dynamo should review the vendor security advisory on GitHub and apply available patches or mitigations immediately. Given the breadth of potential impact across code execution, data integrity, and availability, prioritize remediation for any internet-exposed or multi-tenant Dynamo deployments.

Vendor Advisory

🟠 [HIGH] open-webui/open-webui

5 CVEs | CVSS 3.1: 8.7 | AAS 10.5

  • cpe:2.3:a:open-webui:open-webui:*:*:*:*:*:*:*:* (>= 0.10.0, < 0.11.0)
  • cpe:2.3:a:open-webui:open-webui:*:*:*:*:*:*:*:* (>= 0.8.0, < 0.11.0)
  • cpe:2.3:a:open-webui:open-webui:*:*:*:*:*:*:*:* (>= 0.9.6, < 0.11.0)
  • cpe:2.3:a:open-webui:open-webui:*:*:*:*:*:*:*:* (>= 0.9.0, < 0.11.0)

Open WebUI versions 0.10.0 through prior to 0.11.0 are affected by five vulnerabilities, including multiple rated CVSS 8.7 HIGH. The lead issue involves a stored cross-site scripting flaw where crafted math blocks in chat messages trigger a KaTeX rendering failure, causing the raw input to be inserted as executable HTML, allowing script execution in the browser of anyone viewing the conversation including shared chats.

Teams self-hosting Open WebUI for internal AI chat platforms should upgrade to version 0.11.0 immediately. Given that these flaws can be triggered through normal chat interactions and affect all viewers of shared conversations, any multi-user or externally accessible deployments should be prioritized for patching.

Vendor Advisory

🟠 [HIGH] odysseus-dev/odysseus

1 CVE | CVSS 4.0: 8.7 | AAS 10.1

  • cpe:2.3:a:odysseus-dev:odysseus:*:*:*:*:*:*:*:* (< bf325f6)

Odysseus before commit bf325f6 is affected by one vulnerability (CVE-2026-70619, CVSS 8.7 HIGH) involving missing authorization on endpoint management routes that allows any authenticated non-admin user to overwrite the server-wide embedding backend configuration with an attacker-controlled URL. This enables an attacker to redirect all subsequent embedding operations, including chat and document processing, through a malicious server for data exfiltration or response manipulation.

Teams running Odysseus for AI orchestration should update to a build that includes commit bf325f6 or later immediately. Review the vendor advisory for details, audit endpoint configuration for unauthorized changes, and restrict network-level access to the management interface until the fix is confirmed in place.

Vendor Advisory

🟠 [HIGH] veeam/service_provider_console

2 CVEs | CVSS 4.0: 9.5 | AAS 9.8

  • cpe:2.3:a:veeam:service_provider_console:*:*:*:*:*:*:*:* (>= 7.0, < 8.1.0.21999)

Veeam Service Provider Console is affected by two vulnerabilities, including at least one rated CVSS 9.5 HIGH, allowing an unauthenticated attacker to impersonate a managed agent and obtain that agent’s credentials. This is especially dangerous in managed service provider environments where compromised agent credentials could provide lateral access to customer backup infrastructure, a prime target for ransomware operators.

Service providers and enterprises running Veeam Service Provider Console should apply patches immediately per the guidance in Veeam KB4893. Given the unauthenticated attack vector and the central role this console plays in managing downstream backup agents, treat this as an emergency remediation priority.

Vendor Advisory

🟠 [HIGH] perspective-dev/perspective

2 CVEs | CVSS 4.0: 8.7 | AAS 9.6

  • cpe:2.3:a:perspective-dev:perspective:*:*:*:*:*:*:*:* (>= 5.0.0)

Perspective 5.0.0 is affected by two vulnerabilities, including at least one rated CVSS 8.7 HIGH, involving a path traversal flaw that allows unauthenticated remote attackers to read arbitrary files from the server filesystem using literal directory traversal sequences in HTTP request paths. The impact is compounded by a wildcard CORS configuration that exposes retrieved file contents cross-origin, enabling attackers to exfiltrate sensitive data such as system credentials and application secrets from any origin.

Teams using Perspective for data visualization or analytics dashboards should check for updates from the vendor and restrict network access to affected instances immediately. Review the detailed advisory for all five reported issues, and audit any exposed deployments for signs of unauthorized file access.

Vendor Advisory

🟠 [HIGH] djangoproject/django

1 CVE | CVSS 4.0: 8.7 | AAS 9.6

  • cpe:2.3:a:djangoproject:django:*:*:*:*:*:*:*:* (>= 5.2, < 5.2.17)
  • cpe:2.3:a:djangoproject:django:*:*:*:*:*:*:*:* (>= 6.0, < 6.0.8)

Django versions 5.2 before 5.2.17 and 6.0 before 6.0.8 are affected by one vulnerability (CVE-2026-15307, CVSS 8.7 HIGH) in the GeoDjango spatial lookup handling, where untrusted input passed through the GDALRaster constructor can be exploited. The flaw is reachable through the Django admin changelist query string by staff users with view permission, making it relevant for any deployment using GeoDjango with the admin interface enabled.

Teams running Django applications that use GeoDjango, GeometryField, or RasterField should upgrade to Django 5.2.17 or 6.0.8 immediately. Review the Django security releases page for full details, and audit admin access controls to limit exposure until patches are applied.

Vendor Advisory

🟠 [HIGH] bilin_software_and_informatics_consultancy_inc./humanist_digital_human_resources

2 CVEs | CVSS 3.1: 9.8 | AAS 9.1

  • cpe:2.3:a:bilin_software_and_informatics_consultancy:humanist_digital_human_resources:*:*:*:*:*:*:*:*

Bilin Software HUMANIST Digital Human Resources version 26.0 is affected by two vulnerabilities, including at least one rated CVSS 9.8 HIGH (CVE-2026-14175), involving unrestricted file upload that allows attackers to upload a web shell directly to the server. This provides a straightforward path to full remote code execution and server compromise with minimal attacker effort.

Organizations running HUMANIST Digital Human Resources should upgrade to version 26.1 or later immediately. Given that web shell upload vulnerabilities are trivially exploitable and frequently targeted in the wild, any internet-facing instances should be taken offline or isolated until the update is confirmed in place and web roots are audited for unauthorized files.

Vendor Advisory

🟠 [HIGH] microsoft/microsoft_365_apps_for_enterprise

1 CVE | CVSS 3.1: 8.8 | AAS 9.1

  • cpe:2.3:a:microsoft:microsoft_365_apps_for_enterprise:*:*:*:*:*:*:*:*

Microsoft 365 Apps for Enterprise is affected by one vulnerability (CVE-2026-62870, CVSS 8.8 HIGH) involving a use-after-free condition in Microsoft Office Excel that allows an unauthorized attacker to execute arbitrary code over a network. Exploitation likely requires a user to open a malicious Excel file, making this a significant risk in environments where employees regularly receive spreadsheets via email or file-sharing platforms.

Enterprise security teams should apply the latest Microsoft security updates through standard patch channels immediately. Given the widespread deployment of Microsoft 365 Apps and Excel’s prevalence as a phishing lure delivery format, prioritize this update across all managed endpoints and reinforce user awareness around unsolicited spreadsheet attachments.

Vendor Advisory