| CVE |
Vendors |
Products |
Updated |
CVSS v3.1 |
| OpenClaw before 2026.6.5 could forward Authorization headers during MCP SSE redirects. When the affected feature is enabled and reachable, a lower-trust caller or configured input path could execute or persist actions beyond the caller's intended authorization. Impact depends on the operator's configuration and whether lower-trust input can reach the affected path. |
| OpenClaw versions before 2026.5.28 Bot Framework contains an improper input validation vulnerability that allows lower-trust callers to expose bot tokens and credentials by failing to properly validate serviceUrl parameters. Attackers can supply malicious serviceUrl values through configured input paths to retrieve sensitive authentication data outside the trusted boundary. |
| Incorrect authorization in the XML-RPC API of WebPros Plesk before 18.0.78.4 allows a low-privileged authenticated customer to look up domains they do not own, because ownership is enforced only for certain lookup filters and schema validation is bypassed for legacy protocol versions. This results in cross-tenant disclosure of other tenants' FTP credentials stored in cleartext, which can be leveraged to execute code as another tenant's system user. |
| Microsoft AVML before 0.17.0 could follow a symlink when opening a destination output path on Unix, allowing truncation/overwrite of the symlink target. The destructive effect is performed at open-time via O_TRUNC, and can happen before full input validation completes (“truncation-before-validation”). |
| Dell PowerScale OneFS versions 9.5.0.0 through 9.10.1.7, versions 9.11.0.0 through 9.13.0.2 contains an Insertion of Sensitive Information into Log File vulnerability. A low privileged attacker with local access could potentially exploit this vulnerability, leading to Information disclosure. |
| An issue in Pivotal CRM 6.6.4.08 and systems using patch-ghi-15381-cwe-502-20251225.zip (fixed in Pivotal CRM 6.6.5.10 and Patch_CWE502_20260316.zip) allows a remote attacker to execute arbitrary code via the Pivotal.Engine.Client.Services.Conversion.dll component. NOTE: this issue exists because of an incomplete fix for CVE-2026-39253. |
| The Grav API plugin (getgrav/grav-plugin-api) before 1.0.0-rc.16 accepts JWT access tokens through the ?token= URL query parameter on every API route (JwtAuthenticator::extractBearerToken fallback). Because tokens are embedded in URLs, they are logged verbatim in web server access logs, leaked via the Referer header, stored in browser history, and captured by upstream proxy and CDN logs, exposing valid admin access tokens. A leaked token grants unauthorized API access, including reading configuration and user data, creating admin accounts, modifying system settings, and deleting pages. |
| OpenClaw versions before 2026.5.27 contain a token leakage vulnerability in MS Teams outbound requests that allows lower-trust callers to expose Bot Framework tokens. Attackers can access configured input paths to retrieve credentials that should remain within the trusted boundary. |
| OpenClaw versions before 2026.6.1 contain a credential redaction bypass vulnerability in the trajectory export feature that allows lower-trust callers to access data that should remain within trusted boundaries. Attackers can exploit misconfigured input paths or feature accessibility to expose sensitive credentials and data through the export mechanism. |
| Deserialization of untrusted data in Azure Active Directory allows an unauthorized attacker to deny service over a network. |
| decompress before 4.2.2 allows arbitrary hardlink creation during archive extraction, enabling file read disclosure and file corruption. When processing hardlink entries (type === 'link'), the x.linkname field from the archive is passed directly to fs.link() without validation (index.js line 113). An attacker can craft an archive with a hardlink entry whose linkname is an absolute path to any file on the same filesystem. This creates a hardlink inside the extraction directory that shares the same inode as the target file, enabling both reading and overwriting the original file's content. Hardlinks are limited to files on the same filesystem and cannot target directories. |
| Activepieces is an open source AI workflow automation platform. Prior to 0.82.0, the git-sync feature clones a user-configured Git repository into a temporary directory on the server and then writes flow, table, and connection state into it before pushing back, and two separate weaknesses allowed those writes to escape the intended workspace and land on arbitrary paths on the host filesystem: Git's symbolic-link handling was not disabled on the clone, so an attacker who controlled the remote repository could include symlinks that redirected the writes, and several user-supplied identifiers used to build on-disk paths (the repository slug and the externalId of tables, flows, and connections) were not validated against directory-traversal sequences such as ../. On a self-hosted Enterprise Edition deployment, a user authorized to configure or push to a git-sync repository (holding the WRITE_PROJECT_RELEASE permission) could cause the server to overwrite files anywhere the Activepieces process user can write, which depending on host layout can be leveraged for tampering, denial of service, or remote code execution. This issue is fixed in version 0.82.0. |
| Frogman provides headless PBX control through MCP and HTTP API. Prior to 1.6.2, fm_reset_password in Tools/ResetPassword.php:48-53 returned a plaintext password and fm_add_extension in Tools/AddExtension.php:172 returned a plaintext secret; Frogman.class.php:2207-2211 used auditOutcome to JSON-encode those responses into oc_audit_log.detail, allowing any PERM_READ caller with access to fm_audit_search to recover the stored credentials. This issue is fixed in version 1.6.2. |
| SGLang uses an expert-parallel backup subsystem that exposes a ZeroMQ PULL socket on a routable network interface that does not contain authentication or deserialization safeguards, allowing an attacker to provide a malicious pickle file that results in unauthenticated remote code execution when the feature is enabled and the service is reachable over the network. |
| AWS Bedrock AgentCore Python SDK is an open-source Python library that provides client tools for building AI agents on the Amazon Bedrock AgentCore platform.
Unintended logging of sensitive user content in the OpenTelemetry instrumentation in AWS Bedrock AgentCore Python SDK versions 1.4.8 and 1.5.0 might allow a local authenticated user with access to CloudWatch Logs to access raw user prompts and agent responses containing sensitive data via span attributes. The SDK wrote raw user prompts and complete agent responses into OpenTelemetry span attributes on every invocation without filtering or masking. These spans flow into the customer's aws/spans CloudWatch log group, exposing sensitive content to any principal with log read access.
We recommend you upgrade to version 1.5.1 or later. Users who ran affected versions should also review and purge sensitive content from their aws/spans CloudWatch log groups. |
| dbt-mcp is a Model Context Protocol server for interacting with dbt. Prior to 1.17.1, DbtMCP.call_tool() in src/dbt_mcp/mcp/server.py logged the raw arguments dictionary at INFO level before each tool call and at ERROR level on exceptions, and configure_file_logging() wrote those records to dbt-mcp.log when DBT_MCP_SERVER_FILE_LOGGING=true, preserving sensitive sql_query, vars, and node_selection values in plaintext without automatic rotation or deletion. This issue is fixed in version 1.17.1. |
| Deserialization of untrusted data in Microsoft Office SharePoint allows an unauthorized attacker to execute code over a network. |
| Versions of the package mcp-markdownify-server before 1.0.0 are vulnerable to Files or Directories Accessible to External Parties via the get-markdown-file tool. An attacker can craft a prompt that, once accessed by the MCP host, will allow it to read arbitrary files from the host running the server. |
| NVIDIA TensorRT-LLM for any platform contains a vulnerability in visual gen server, where an attacker could cause an unsafe deserialization by unauthorized zeroMQ deserialization. A successful exploit of this vulnerability might lead to code execution. |
| An insecure symlink following vulnerability exists in Canonical ubuntu-pro-client (formerly ubuntu-advantage-tools) within the pro collect-logs command framework. The utility creates or utilizes predictable temporary file paths or user-accessible log directories when gathering diagnostic information without verifying the file type or ownership. An unprivileged local attacker can exploit this behavior by creating a symbolic link (symlink) at a predictable destination path pointing to an arbitrary, root-readable file (such as /etc/shadow or private files within /root). When a root administrator or operator subsequently executes the pro collect-logs command, the tool follows the user-controlled symlink, reads the target file, and compresses its contents into the resulting diagnostic support archive. Because the output archive remains readable by the unprivileged user, the attacker can extract and read the sensitive root-owned files, leading to a complete information disclosure of system secrets. |
