10 vulnerabilities across 8 products scored HIGH or above on August 10, 2026.

  • 🔴 CRITICAL: 2
  • 🟠 HIGH: 8

Exploit Status Upgrades

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

  • [UPGRADED] CVE-2026-71558 (apache/fory) — F1: exploitable → functional, AAS: 10.9 → 12.9 (HIGH → CRITICAL). Originally in 2026-08-07 bulletin.
  • [UPGRADED] CVE-2026-62873 (microsoft/microsoft_365_admin_center) — F1: theoretical → poc, AAS: 9.1 → 11.6 (HIGH → HIGH). Originally in 2026-08-07 bulletin.
  • [UPGRADED] CVE-2026-65507 (sergey/aiwu) — F1: exploitable → functional, AAS: 10.9 → 12.9 (HIGH → CRITICAL). Originally in 2026-08-06 bulletin.
  • [UPGRADED] CVE-2026-5430 (wso2/wso2_universal_gateway) — F1: theoretical → poc, AAS: 10.7 → 13.2 (HIGH → CRITICAL). Originally in 2026-08-06 bulletin.
  • [UPGRADED] CVE-2026-34191 (apache/apr-util) — F1: exploitable → functional, AAS: 10.5 → 13.7 (HIGH → CRITICAL). Originally in 2026-08-06 bulletin.
  • [UPGRADED] CVE-2026-28139 (wpdreams/ajax_search_lite) — F1: exploitable → functional, AAS: 10.1 → 12.1 (HIGH → CRITICAL). Originally in 2026-08-06 bulletin.
  • [UPGRADED] CVE-2025-15039 (wso2/wso2_identity_server) — F1: theoretical → itw, AAS: 9.9 → 13.9 (HIGH → CRITICAL). Originally in 2026-08-06 bulletin.
  • [UPGRADED] CVE-2026-65583 (apache/cxf) — F1: exploitable → functional, AAS: 9.9 → 11.9 (HIGH → HIGH). Originally in 2026-08-06 bulletin.
  • [UPGRADED] CVE-2026-68079 (apache/cxf) — F1: theoretical → poc, AAS: 9.6 → 12.1 (HIGH → CRITICAL). Originally in 2026-08-06 bulletin.
  • [UPGRADED] CVE-2026-48168 (mervinpraison/praisonai) — F1: exploitable → functional, AAS: 10.8 → 12.8 (HIGH → CRITICAL). Originally in 2026-08-05 bulletin.
  • [UPGRADED] CVE-2026-70478 (flowiseai/flowise) — F1: theoretical → poc, AAS: 9.9 → 12.4 (HIGH → CRITICAL). Originally in 2026-08-04 bulletin.
  • [UPGRADED] CVE-2026-39931 (openemr/openemr) — F1: exploitable → functional, AAS: 9.1 → 11.1 (HIGH → HIGH). Originally in 2026-08-03 bulletin.
  • [UPGRADED] CVE-2026-38447 (osticket/osticket) — F1: theoretical → poc, AAS: 9.6 → 12.1 (HIGH → CRITICAL). Originally in 2026-08-03 bulletin.
  • [UPGRADED] CVE-2026-46713 (misskey-dev/misskey) — F1: theoretical → poc, AAS: 9.1 → 11.6 (HIGH → HIGH). Originally in 2026-08-03 bulletin.

🔴 [CRITICAL] alseambusher/crontab-ui

2 CVEs | CVSS 3.1: 9.8 | AAS 12.1

  • cpe:2.3:a:alseambusher:crontab-ui:*:*:*:*:*:*:*:* (< 0.4.3)

Crontab UI by alseambusher, versions through 0.4.2, is affected by 2 critical vulnerabilities including at least one OS command injection flaw rated CVSS 9.8. An unauthenticated remote attacker can execute arbitrary system commands by importing a crafted crontab database file through the POST /import endpoint, which accepts and processes .db files without validation or sanitization. Proof-of-concept exploit code is available, increasing the likelihood of active exploitation.

Organizations running crontab-ui for managing cron jobs on Linux servers should treat this as an urgent priority. There is currently no patched release available. Administrators should immediately restrict network access to crontab-ui instances, place them behind authentication if not already protected, and monitor the vendor’s GitHub repository at github.com/alseambusher/crontab-ui for a security update. Consider removing crontab-ui from production environments until a fix is released.

Vendor Advisory

🟠 [HIGH] tencent/apijson

1 CVE | CVSS 3.1: 9.8 | AAS 11.1

  • cpe:2.3:a:tencent:apijson:*:*:*:*:*:*:*:*

Tencent APIJSON through version 8.1.8 is affected by a critical SQL injection vulnerability (CVSS 9.8) that allows unauthenticated remote attackers to bypass per-table access controls and read arbitrary database tables. The flaw exists in the Map-form @having operator, where the APIJSONORM library fails to apply role-based allow-list checks, enabling injection of raw SQL fragments to access tables that should be restricted to no roles.

Organizations using APIJSON as a backend API framework should treat this as high priority, as the vulnerability requires no authentication and is confirmed exploitable. Administrators should monitor the vendor’s GitHub repository at github.com/Tencent/APIJSON for a patched release and apply it immediately when available. In the interim, consider restricting network access to APIJSON endpoints, deploying a web application firewall to filter malicious @having parameters, and auditing database logs for signs of unauthorized table access.

Vendor Advisory

🟠 [HIGH] asyncfuncai/deepwiki-open

1 CVE | CVSS 3.1: 9.8 | AAS 11.1

  • cpe:2.3:a:asyncfuncai:deepwiki-open:*:*:*:*:*:*:*:*

AsyncFuncAI deepwiki-open through commit 16f35a0 is affected by a critical path traversal vulnerability (CVSS 9.8) that allows unauthenticated remote attackers to write to or delete arbitrary files with root privileges. The wiki-cache API endpoint constructs file paths from user-controlled input without sanitization, and when combined with a second unauthenticated directory creation endpoint, provides full arbitrary file write and delete primitives on the host system.

Teams running deepwiki-open instances should treat this as an urgent priority given the unauthenticated access, root-level impact, and confirmed exploitability. Monitor the vendor’s GitHub repository at github.com/AsyncFuncAI/deepwiki-open for a fix and apply it immediately. Until patched, restrict all network access to deepwiki-open instances or take them offline entirely, as no authentication barrier exists to slow exploitation.

Vendor Advisory

🟠 [HIGH] lobehub/lobe-chat

1 CVE | CVSS 3.1: 7.6 | AAS 10.9

  • cpe:2.3:a:lobehub:lobe-chat:*:*:*:*:*:*:*:* (< 2.2.14)

LobeHub Lobe-Chat through version 2.2.13 is affected by a stored cross-site scripting vulnerability (CVSS 7.6) that allows a low-privileged authenticated user to execute arbitrary JavaScript in the browsers of other users. The flaw exists in the avatar upload handler, which derives the stored file extension and content type from the original filename rather than inspecting file content, enabling an attacker to upload a crafted SVG file containing malicious script elements.

Organizations deploying Lobe-Chat should be aware that proof-of-concept exploit code is available and any authenticated user can weaponize this to target other users, including administrators. Monitor the vendor’s GitHub repository at github.com/lobehub/lobe-chat for a patched release and apply it promptly. As an interim mitigation, consider restricting avatar upload file types at the reverse proxy or storage layer to block SVG files, and review existing uploaded avatars for embedded script content.

Vendor Advisory

🟠 [HIGH] dulldusk/phpfm

2 CVEs | CVSS 3.1: 9.8 | AAS 10.6

  • cpe:2.3:a:dulldusk:phpfm:*:*:*:*:*:*:*:* (< 1.8.1)

Dulldusk phpfm through version 1.8.0 is affected by 2 critical vulnerabilities (CVSS 9.8) including at least one unrestricted file upload flaw that allows unauthenticated remote attackers to execute arbitrary PHP code on the server. The application ships with no authentication enabled and an empty upload extension filter by default, allowing any remote attacker to upload and execute a PHP webshell with no barriers.

Anyone running phpfm on a web-accessible server should treat this as an emergency, as proof-of-concept exploit code is available and the default configuration is immediately exploitable without any credentials. Take phpfm instances offline immediately or restrict access to trusted networks only. Monitor the vendor’s GitHub repository at github.com/dulldusk/phpfm for a security update, and audit any exposed instances for signs of webshell uploads or unauthorized file modifications.

Vendor Advisory

🟠 [HIGH] fosrl/pangolin

1 CVE | CVSS 3.1: 9.6 | AAS 10.4

  • cpe:2.3:a:fosrl:pangolin:*:*:*:*:*:*:*:*

Fosrl Pangolin through version 1.20.0 is affected by a critical improper authorization vulnerability (CVSS 9.6) that allows an authenticated remote attacker to access any resource in any organization by reusing an access token issued for a different resource. The token verification handler fails to validate the target resource ID, meaning a token scoped to one resource passes verification for all other resources across all organizations.

Any organization using Pangolin for resource access control should treat this as high priority, as the flaw is confirmed exploitable and completely defeats cross-resource and cross-organization access boundaries. Monitor the vendor’s GitHub repository at github.com/fosrl/pangolin for a patched release and apply it immediately. Until a fix is available, audit access logs for signs of token reuse across unexpected resources and consider implementing additional network-layer access restrictions to limit exposure.

Vendor Advisory

🟠 [HIGH] daptin/daptin

1 CVE | CVSS 3.1: 9.1 | AAS 9.9

  • cpe:2.3:a:daptin:daptin:*:*:*:*:*:*:*:*

Daptin through version 0.12.34 is affected by a critical improper authorization vulnerability (CVSS 9.1) that allows unauthenticated remote attackers to read, create, update, and delete usergroup records. The permission check functions fail to reject null or zero user ID references, meaning unauthenticated requests pass authorization checks on any table rows that lack an associated user account, including all usergroup records.

Organizations using Daptin as a backend API framework should treat this as high priority, as the vulnerability is confirmed exploitable and grants full CRUD access to group membership without any credentials. Monitor the vendor’s GitHub repository at github.com/daptin/daptin for a patched release and apply it as soon as available. In the interim, restrict network access to Daptin instances to trusted sources only and audit usergroup records for unauthorized modifications.

Vendor Advisory

🟠 [HIGH] duhow/xiaoai-patch

1 CVE | CVSS 3.1: 9.8 | AAS 9.1

  • cpe:2.3:a:duhow:xiaoai-patch:*:*:*:*:*:*:*:*

Duhow xiaoai-patch through commit fb07049 is affected by a critical OS command injection vulnerability (CVSS 9.8) that allows remote attackers to execute arbitrary system commands on Xiaomi smart speakers running the patch. The /mute and /unmute API endpoints pass user-supplied input directly to os.system() without any sanitization, enabling trivial command injection via shell metacharacters.

Users who have installed xiaoai-patch on Xiaomi smart speakers should treat this as an urgent issue, as the vulnerability is confirmed exploitable and provides full system command execution with no authentication required. Disable or remove xiaoai-patch immediately and restrict network access to any patched devices until a fix is available. Monitor the vendor’s GitHub repository at github.com/duhow/xiaoai-patch for a security update, and audit affected speakers for signs of unauthorized command execution or compromise.

Vendor Advisory