Trigger.dev has a vulnerability where the deployment cancel operation is scoped only by project ID and not by environment ID. This allows an API key for a lower-trust environment (e.g., dev) to cancel deployments in higher-trust environments (e.g., production) within the same project. The deployment GET route correctly enforces environment scoping, but the cancel route does not, creating an authorization bypass. This affects both self-hosted and cloud deployments of Trigger.dev prior to version 4.5.6.
A cross-tenant SQL injection vulnerability exists in trigger.dev's TSQL query compiler at the POST /api/v1/query endpoint. The vulnerability arises because the window-function name in the query is concatenated into the SQL string without allowlisting or escaping, allowing an authenticated customer to inject raw subqueries that bypass tenant isolation. This enables an attacker to read analytics data belonging to other tenants. The issue has been verified end-to-end against the real compiler and a live ClickHouse database.
Trigger.dev versions prior to 4.5.5 have an unauthenticated POST endpoint allowing arbitrary data injection into realtime streams by anyone who knows or can guess a run friendlyId. The endpoint lacks authentication and environment scoping, enabling cross-environment access. This vulnerability has a medium severity with a CVSS score of 5.3.
Self-hosted trigger.dev version 4 instances deployed with the default Docker Compose configuration using hardcoded secrets in the provided .env.example file are vulnerable to unauthenticated attacks. The MAGIC_LINK_SECRET allows attackers to forge authentication tokens and gain access as any user without email verification. This leads to automatic user creation and full access to infrastructure services including the database, Redis, and ClickHouse, all of which use default or no authentication. The vulnerability enables complete infrastructure compromise, including tenant data, API keys, encrypted secrets, and job queue manipulation.
Trigger.dev versions prior to 4.5.4 contain multiple critical vulnerabilities in the self-hosted coordinator Socket.IO namespace. A hardcoded default secret allows unauthenticated access, enabling attackers to retrieve decrypted environment variables from any run if they know the run's internal ID. Additionally, unauthenticated write handlers allow attackers to send forged messages affecting arbitrary runs. These issues were fully resolved by removing the vulnerable V1 execution stack and default secrets in version 4.5.4. Managed cloud deployments were not affected due to non-default secrets.
Xray-core has a vulnerability where pinning a CA certificate using pinnedPeerCertSha256 can enable man-in-the-middle (MITM) attacks if the ServerName configuration is empty. This occurs because the TLS verification does not check the hostname against the certificate when ServerName is missing and InsecureSkipVerify is true. In some code paths, ServerName is not set when using IP addresses or certain dialers, allowing an attacker to issue a leaf certificate from the pinned CA and hijack connections.
Trigger.dev contains a blind Server-Side Request Forgery (SSRF) vulnerability in its webhook alert channel implementation. Authenticated tenants can configure webhook URLs without validation, allowing the server to POST alert payloads to internal or metadata service IP addresses. This can lead to internal host scanning, state-changing POST requests to internal services, and potential access to IMDSv1 via redirects. The vulnerability affects versions prior to 4.5.2. A fix is available that includes validating webhook URLs against private, link-local, loopback, and metadata IP ranges and re-checking URLs at fetch time.
Trigger.dev contains an insecure direct object reference (IDOR) vulnerability in the run replay action function where no authentication or authorization checks are performed. This allows any authenticated user to replay task runs from any organization by knowing the run's friendlyId. A similar issue exists in the batch completion action, enabling unauthorized triggering of batch completions. The vulnerability enables cross-organization task execution with victim environment secrets, resource consumption, and potential data integrity issues. Patch is available.
The web management service in affected RouterOS versions contains an integer underflow in its HTTP request body handling that is reachable before authentication. This can be leveraged by an unauthenticated network attacker to achieve arbitrary code execution as root, or to cause a denial of service, using a single crafted request.
UTMStack versions prior to 11.2.16 have an account enumeration vulnerability. This flaw allows unauthenticated attackers to determine if an email address is registered by submitting it to the password reset initiation endpoint. Registered emails return a 200 OK response, while unregistered ones cause a 500 Internal Server Error with backend details. This behavior enables attackers to distinguish valid accounts, potentially facilitating targeted phishing or credential attacks.
UTMStack versions prior to 11.2.16 have a JPQL injection vulnerability in the UtmNetworkScanService.searchPropertyValues() method. Authenticated attackers can exploit this flaw by injecting malicious JPQL through the value parameter in the GET /api/utm-network-scans/searchPropertyValues endpoint. This allows unauthorized reading of arbitrary entity data, including sensitive credential tables such as jhi_user.
UTMStack versions prior to 11.2.16 contain a server-side request forgery (SSRF) vulnerability. Authenticated attackers can exploit this flaw by supplying an unvalidated URL parameter to the PdfService.downloadPdf() method via the GET /api/generate-pdf-report endpoint. This allows the attacker to make the server request arbitrary internal resources, potentially exposing sensitive internal data rendered into the generated PDF report.
UTMStack versions before 11.2.16 have an authentication bypass vulnerability (CVE-2026-82042) that allows remote attackers to gain full administrative API access by using a valid Utm-Internal-Key header matching the INTERNAL_KEY environment variable. This bypass does not require a user account or JWT and lacks path restrictions, constant-time comparison, rate limiting, or audit logging. Exploitation enables attackers to create accounts, manage users, exfiltrate data, and modify security rules.
UTMStack versions before 11.2.16 have a missing authorization vulnerability in the UTMIncidentCommandWebsocket.processCommand() handler. This handler, mapped to the /command/{hostname} STOMP destination, does not enforce role checks or command allowlists before forwarding commands. As a result, any authenticated user can execute arbitrary operating-system commands on monitored endpoints where agents typically run with high privileges.
UTMStack versions before 11.2.16 have a SQL injection vulnerability in the UtmAssetGroupService.searchQueryBuilder() method. Authenticated attackers can exploit this flaw by sending crafted requests to the GET /api/utm-asset-groups/searchGroupsByFilter endpoint, injecting arbitrary SQL commands. This vulnerability allows execution of SQL with DBA privileges, potentially leading to full database read, data modification, and filesystem access.
A Cross-Site Request Forgery (CSRF) vulnerability exists in Webriti Wallstreet up to version 2.8.6. This vulnerability allows an attacker to trick a user into submitting unauthorized requests to the application, potentially leading to full compromise of confidentiality, integrity, and availability.
UTMStack versions prior to 11.2.16 have a server-side request forgery (SSRF) vulnerability in the IdentityProviderService.validateMetadataUrl() function. This flaw allows authenticated attackers to supply a malicious metadata URL to the identity-providers endpoint, causing the server to send requests to arbitrary internal or cloud metadata hosts. The vulnerability affects the POST/PUT /api/identity-providers endpoint, which does not validate the target host, IP address, or URL scheme. Exploitation can enable internal network port scanning and access to cloud instance-metadata services.
The All in One SEO WordPress plugin versions before 5.0.2.1 contain a vulnerability where the plugin fails to correctly identify shortcodes in user-derived content before stripping them. This allows unauthenticated users to execute arbitrary shortcodes registered on the site. On sites upgraded from older versions, the protection mechanism is completely disabled, increasing the risk of exploitation.
The voxeval package on PyPI versions 0.4.2 through 0.4.5 contains malicious code that deploys a cryptocurrency miner instead of providing legitimate functionality. This is a malicious campaign identified as 2026-10-voxeval with clear intent to misuse victim resources for cryptomining.
The npm package 'ui-base-colors' contains malicious code that fully compromises any computer on which it is installed or running. This compromise includes potential unauthorized access to all secrets and keys stored on the affected system. Removal of the package alone does not guarantee elimination of all malicious software introduced by it.