A code injection vulnerability (CWE-94) exists in heymrun's heym product prior to version 0.0.75. The vulnerability occurs in the POST /api/mcp/fetch-tools endpoint, which allows any authenticated user with the lowest role to supply commands and arguments that are executed on the shared backend host before protocol validation. This can lead to high impact including full confidentiality, integrity, and availability compromise. The issue has been patched in version 0.0.75.
The URL http://vagos.cc/Sakura.sh was observed in attacker activity by OffSeq Mirage honeypot sensors on 2026-11-11. It is associated with the MITRE ATT&CK technique T1105 (Ingress Tool Transfer), indicating it is used for downloading attacker tools or payloads. The attacker infrastructure was identified in the Netherlands. This URL has been seen in attacker payloads such as tool downloads, callbacks, or exploitation targets. Organizations are advised to block and monitor this URL and search their proxy, web, or WAF logs for any requests to it.
The URL http://botnet.code.wf/Sakura.sh was observed twice in attacker activity on 2026-11-11 by OffSeq Mirage honeypot sensors. It is associated with the ATT&CK technique T1105 (Ingress Tool Transfer) and originates from attacker infrastructure in the Netherlands. This URL is used in attacker payloads for tool download, callback, or exploitation targeting. Organizations are advised to block and monitor this URL and review proxy, web, and WAF logs for any requests to it.
CVE-2026-67545 is a server-side request forgery (SSRF) vulnerability in the HTTP workflow node of the Heym platform prior to version 0.0.73. The HTTP node makes outbound requests to user-supplied URLs without validating the target host or IP address. This allows an unauthenticated attacker to trigger SSRF if a workflow is exposed with anonymous execution and targets internal or private network addresses. The vulnerability has been fixed in version 0.0.73.
CVE-2026-67544 is a code injection vulnerability in the Heym workflow engine prior to version 0.0.71. The vulnerability allows any authenticated user, including newly registered accounts, to execute arbitrary Python code on the backend host by supplying malicious code in the playwrightCode field. The code is executed without sandboxing or validation, leading to full remote code execution with the privileges of the service process user, which is root in the default Docker deployment. This issue has been fixed in version 0.0.71.
CVE-2026-67543 is a code injection vulnerability in the heymrun heym product prior to version 0.0.70. The vulnerability arises because the skill executor runs untrusted Python code without Docker sandbox isolation, allowing execution with filesystem access, network egress, and environment variables exposure. In the default Docker configuration, this can escalate to host root privileges. The issue has been fixed in version 0.0.70.
A Server-Side Request Forgery (SSRF) vulnerability exists in heymrun heym versions prior to 0.0.70. Authenticated users with the lowest role can access the POST /api/mcp/fetch-tools endpoint, which fetches attacker-supplied URLs without proper validation or allowlisting. The vulnerability allows attackers to control outbound request headers and gain a semi-blind SSRF oracle via error reflection in HTTP 422 responses. This issue is fixed in version 0.0.70.
A privilege assignment vulnerability in heymrun heym prior to version 0.0.70 allows any non-creator team member to add attacker accounts to a team. This occurs because team-member management is controlled only by team membership without differentiating roles, enabling unauthorized users to gain access to all credentials shared within the team. The issue is fixed in version 0.0.70.
A vulnerability in heym prior to version 0.0.70 allows any authenticated user to access private workflow templates belonging to other tenants. The get_workflow_template and get_node_template functions do not enforce visibility restrictions on single-item fetches, exposing sensitive template data including node/edge graphs and shared email addresses. This authorization bypass has been fixed in version 0.0.70.
CVE-2026-126229 is a critical improper access control vulnerability in Fortinet FortiManager and FortiAnalyzer products, including their cloud versions. It affects multiple versions from 7.0.0 through 8.0.1. The flaw allows an authenticated administrator to bypass authorization controls and access other Virtual Domains (VDOMs) via specially crafted HTTP requests. This vulnerability has a high CVSS score of 9.1, indicating severe confidentiality, integrity, and availability impacts. An official fix is available for the affected versions.
Roundcube Webmail versions before 1.6.21 and 1.7.x before 1.7.6 have an IMAP command injection vulnerability. This flaw allows attackers to inject IMAP commands via CRLF sequences in the _uid parameter when the messages_cache feature is enabled. Exploitation requires a logged-in victim to open a specially crafted GET link, which can lead to unauthorized deletion of messages, modification of message flags, or unauthorized mailbox access via SETACL commands.
blueprintUE Self-Hosted Edition versions prior to 4.3.6 have an iframe injection vulnerability. Authenticated users can exploit this by embedding arbitrary third-party content via the PeerTube video URL matcher. This allows loading attacker-controlled content in an unsandboxed iframe after user consent, potentially framing fake login forms or malware downloads. The vulnerability has a low CVSS score of 3.5 and requires user interaction.
WWBN AVideo versions 14.3 through 29.7.1 are affected by a stored cross-site scripting (XSS) vulnerability. This flaw allows uploaders to inject malicious scripts via the videoDownloadedLink parameter, which is stored without proper escaping. The injected scripts execute in administrator sessions through a Source tag rendered in the video management interface, potentially enabling attackers to install plugin archives containing PHP code.
WWBN AVideo versions up to 29.7.1 have a missing authorization vulnerability in the objects/import.json.php endpoint. This flaw allows users with uploader privileges to import MP4 files from other users by referencing any folder under videos/. Exploitation enables attackers to copy private, unlisted, password-protected, or pay-per-view videos, read associated description files (.txt/.html), list files via objects/listFiles.json.php, and delete media using the delete parameter.
WWBN AVideo versions 29.1.0 through 29.7.1 have a stored cross-site scripting (XSS) vulnerability in the channel About field. Registered users can inject malicious markup that bypasses filtering and executes when the channel feed is viewed, potentially affecting any visitor including administrators.
weDevs ERP (WP ERP) versions through 1.17.11 have an incorrect authorization vulnerability that allows authenticated employees to modify their own HR records. This occurs because HRM AJAX handlers check the erp_edit_employee capability instead of the intended erp_manage_jobinfo capability. Exploiting this flaw, attackers can change sensitive job-related information such as designation, reporting manager, pay rate, employment type, or delete job history.
BookStack versions prior to 26.09.3 have an authorization vulnerability that allows authenticated users to access or manipulate other users' unpublished draft page images. The flaw arises because the system only checks PageView permission without verifying draft status, enabling attackers to read or tamper with draft images via sequential IDs.
WuzAPI versions up to 1.1.0 have a vulnerability in the GetHistory handler that allows authenticated users to access chat metadata from other tenants. This cross-tenant information disclosure lets attackers with message history enabled retrieve internal user IDs, stored chat JIDs, and last-message timestamps by making specific GET requests.
blueprintUE Self-Hosted Edition versions prior to 4.3.7 have an insecure default Docker Compose configuration that sets MYSQL_ALLOW_EMPTY_PASSWORD=yes and exposes the MariaDB database on port 3306. This allows unauthenticated attackers with network access to the Docker host's port 3306 to log in as root without a password, enabling them to read, modify, or delete all user accounts, blueprints, and databases.
WWBN AVideo versions through 29.7.1 have an incorrect authorization vulnerability that allows delegated Videos Admin users to retrieve video_id_hash values of any video owner via encoder endpoints. This flaw enables attackers to obtain hashes for administrator-owned videos and use them to perform password-less logins, gaining unauthorized access to plugin upload, configuration, and user management features.