CVE Database

Devika v1.0 is vulnerable to Code Injection via the Runner.run_code function in src/agents/runner/runner.py.

Devika v1.0 is vulnerable to Code Injection in the Runner.execute function in src/agents/runner/runner.py which allows an attacker to achieve arbitrary code execution by exploiting the direct execution of LLM-generated content.

DeepTutor v1.4.0 is vulnerable to command execution in /tutorbot/agent/tools/shell.py:ExecTool.execute.

agentgpt v.1.0.0 is vulnerable to Incorrect Access Control in next/src/server/api/routers/agentRouter.ts. An externally reachable path accepts a caller-selected object or tenant identifier and reaches a data-access operation without a visible owner, tenant, workspace, or membership binding on that object.

In DB-GPT 0.7.5 and 0.8.0, a skill uploaded through the real /api/v1/skills/upload route can later be executed through the real /api/v1/chat/react-agent flow.

DB-GPT v0.7.5 and v0.8.0 contains directory traversal in python_file_upload (packages/dbgpt-app/src/dbgpt_app/openapi/api_v1/python_upload_api.py:42). A remote attacker can use the validated exploitation path to write files outside the intended workspace or storage boundary.

bisheng 2.3.0, 2.4.0, and 2.4.0-beta1 is vulnerable to Code Injection in src/backend/bisheng/api/v1/validate.py.

bisheng 2.3.0, 2.4.0, and 2.4.0-beta1 is vulnerable to directory traversal in save_download_file (src/backend/bisheng/core/cache/utils.py:290).

In camel-ai camel 0.2.91a1, v0.2.91a2 and v0.2.91a3, TerminalToolkit.shell_exec allows prompt-driven shell command execution without an approval boundary.

In camel-ai camel 0.2.91a1, v0.2.91a2 and v0.2.91a3, CodeExecutionToolkit can run model-produced Python code through SubprocessInterpreter without an approval boundary.

In agentscope 1.0.18, 1.0.19, and 1.0.19 when the RealtimeAgent session exposes execute_python_code as an available tool, a remote WebSocket user can prompt the agent to call that tool and run Python code in the service environment. In the validated path, RealtimeAgent._acting forwards the model-produced tool call to Toolkit.call_tool_function, which invokes execute_python_code without an additional approval or isolation boundary on that path.

Kiteworks Email Protection Gateway did not sufficiently restrict which account a certificate could be assigned to. This could allow an attacker to associate a certificate with another user's account, affecting the confidentiality and integrity of that account's encrypted mail and, where certificate-based login is enabled, potentially permitting unauthorized access to the account.

A stored cross-site scripting (XSS) weakness in Kiteworks Core could allow an unauthenticated attacker to store crafted content that later executes arbitrary JavaScript in the authenticated session of an administrator who views the affected page. This could have permitted the attacker to gain full administrative control, including the creation of a new administrative account.

Kiteworks Core did not correctly validate a parameter submitted to the password reset workflow. An unauthenticated attacker who knew the email address of a user with a locally stored password could potentially reset that account's password without access to the emailed reset link and then authenticate as that user, including where the account holds administrative privileges.

Improper authentication in a Kiteworks Email Protection Gateway administrative service. An administrative service in Kiteworks Email Protection Gateway did not consistently enforce administrator authentication, so the required password check could be bypassed. An attacker who referenced a valid administrator account could potentially create, modify, or delete internal users and managed domains and change their security-feature configuration without authenticating; deleting a managed domain also removes its user accounts and could lock administrators out of the gateway.

Kiteworks Email Protection Gateway before version 9.5.0 is vulnerable to Server-Side Request Forgery (SSRF). A server-side request forgery (SSRF) weakness in Kiteworks Email Protection Gateway could allow a remote, unauthenticated attacker to induce the gateway to issue crafted requests to internal or otherwise unintended network destinations. The requests are triggered while the gateway renders message content that references external resources. Depending on the services reachable from the gateway, this could disclose sensitive internal information or trigger unintended actions on internal systems.

Kiteworks Email Protection Gateway before version 9.5.0 is vulnerable to Server-Side Request Forgery (SSRF). A server-side request forgery (SSRF) weakness in Kiteworks Email Protection Gateway could allow a remote, unauthenticated attacker to induce the gateway to issue crafted requests to internal or otherwise unintended network destinations. The requests are triggered while the gateway performs an online certificate status check for an inbound message. Depending on the services reachable from the gateway, this could disclose sensitive internal information or disrupt gateway operation.

Kiteworks Email Protection Gateway before version 9.5.0 is vulnerable to Server-Side Request Forgery (SSRF). A server-side request forgery (SSRF) weakness in Kiteworks Email Protection Gateway could allow a remote, unauthenticated attacker to induce the gateway to issue crafted requests to internal or otherwise unintended network destinations. The requests are triggered while the gateway retrieves a certificate revocation list in an inbound message. Depending on the services reachable from the gateway, this could disclose sensitive internal information or disrupt gateway operation.

Kiteworks Email Protection Gateway before version 9.5.0 is vulnerable to Server-Side Request Forgery (SSRF). A server-side request forgery (SSRF) weakness in Kiteworks Email Protection Gateway could allow a remote, unauthenticated attacker to induce the gateway to issue crafted requests to internal or otherwise unintended network destinations. The requests are triggered while the gateway retrieves an issuer certificate in an inbound message. Depending on the services reachable from the gateway, this could disclose sensitive internal information or disrupt gateway operation.

Kiteworks Email Protection Gateway before version 9.5.0 is vulnerable to Server-Side Request Forgery. Kiteworks Email Protection Gateway performed server-side fetches of URLs contained in the message content it processed, without adequately restricting the fetch destination. A remote, unauthenticated sender could craft a message that caused the gateway to issue requests to internal services and cloud instance metadata endpoints and return the responses, potentially disclosing sensitive internal data and, depending on the internal service reached, affecting its state.

iperf3 3.21 (esnet/iperf) contains a remote, unauthenticated heap use-after-free: the server's per-test watchdog server_timer_proc() frees streams without cancelling/joining their worker threads, so a blocked worker dereferences a freed iperf_stream; fixed in 3.22.

In ldapd in OpenBSD 7.8 before errata 057 and 7.9 before errata 021, delegated BSD authentication results are correlated only by the LDAP child process client file descriptor and LDAP message ID. After a connection closes, a later connection that reuses the same file descriptor and message ID can receive the earlier authentication result. A remote attacker who can reach ldapd can complete a Bind as another identity. A missing connection can also cause a NULL pointer dereference. (ldapd is not enabled by default.)

piscina is a node.js worker pool implementation. Prior to 4.9.4, 5.3.2, and 6.0.0-rc.5, Piscina stores ThreadPool.options in src/index.ts as a plain object that inherits from Object.prototype. Applications with a separate prototype-pollution primitive can therefore supply inherited values for security-sensitive options that do not have own defaults. An inherited execArgv value is passed to the Node.js Worker constructor and can preload attacker-controlled code in worker threads, an inherited loadBalancer function can execute during task scheduling, and inherited env values can alter worker environments. This issue is fixed in versions 4.9.4, 5.3.2, and 6.0.0-rc.5.

Kobako is a Ruby gem that embeds a Wasm-isolated mruby interpreter inside applications, allowing execution of untrusted Ruby scripts (LLM-generated code, user formulas, student submissions, third-party plugins) in-process without giving them access to host memory, files, network, or credentials. From version 0.1.0 to before version 0.9.1, a guest mruby script running inside the Kobako sandbox can execute arbitrary Ruby in the host process, fully escaping the sandbox. This issue has been patched in version 0.9.1.

yii2-starter-kit through 4.2.0 exposes the Yii debug and Gii modules to all IP addresses by setting allowedIPs to ['*'] in its default development configuration. Unauthenticated remote attackers can access the debug endpoint to read sensitive data including session cookies and database queries, or access the Gii endpoint to generate and write PHP files into the application directory.

Contributor PHP Object Injection in Nested Pages <= 3.3.2 versions.

Missing authentication for critical function vulnerability for all PTZOptics cameras and the Firmware Upgrade Tool - Firmware Update modules. A missing authentication vulnerability in the firmware update mechanism of affected PTZOptics cameras allows an unauthenticated user to install modified firmware on the device without administrator credentials. This vulnerability allows attackers to upload modified firmware to the device without admin credentials. This issue affects: * Move 4K 12X before: 0.0.98 * Move 4K 20X before: 0.1.33 * Move 4K 30X before: 2.1.17 * Link 4K 12X before: 0.0.99 * Link 4K 20X before: 0.1.37 * Link 4K 30X before: 2.1.18 * Move SE 12X before: 9.1.66 * Move SE 20X before: 9.1.44 * Move SE 30X before: 9.1.46 * Studio 4K 12X before: 8.3.32 * Studio 4K 20X before: 8.3.32 * Studio SE 12X before: 8.3.32 * Studio SE 20X before: 8.3.32 * All Generation 2 cameras, including: PT12X-SDI-GY-G2, PT12X-SDI-WH-G2, PT12X-NDI-GY-G2, PT12X-NDI-WH-G2; PT12X-USB-GY-G2, PT12X-USB-WH-G2; PT20X-SDI-GY-G2, PT20X-SDI-WH-G2, PT20X-NDI-GY-G2, PT20X-NDI-WH-G2; PT20X-USB-GY-G2, PT20X-USB-WH-G2; PT30X-SDI-GY-G2, PT30X-SDI-WH-G2, PT30X-NDI-GY-G2, PT30X-NDI-WH-G2; PTVL-ZCAM, PTVL-NDI-ZCAM; PTEPTZ-ZCAM-G2, PTEPTZ-NDI-ZCAM-G2; PT12X-ZCAM, PT12X-NDI-ZCAM; PT20X-ZCAM, PT20X-NDI-ZCAM; Studio Pro - All versions * Upgrade Tool - All versions

Tugtainer is a self-hosted app for automating updates of Docker containers. Prior to version 1.30.6, Tugtainer allows an authenticated user to make the backend server send outbound HTTP requests to arbitrary user-supplied URLs through the notification test endpoint. The /settings/test_notification endpoint accepts a urls field and passes it directly to Apprise without restricting protocols, hostnames, localhost addresses, private IP ranges, or cloud metadata addresses. This can be abused as an authenticated blind server-side request forgery (SSRF). This issue has been patched in version 1.30.6.

Tugtainer is a self-hosted app for automating updates of Docker containers. Prior to version 1.30.4, Tugtainer Agent allows unauthenticated access to Docker management APIs when AGENT_SECRET is not configured. The Agent uses request signatures to protect its API routes. However, in agent/auth.py, the signature verification function returns successfully if Config.AGENT_SECRET is empty. This causes protected Agent APIs to become accessible without authentication. This issue has been patched in version 1.30.4.

Tugtainer is a self-hosted app for automating updates of Docker containers. Prior to version 1.30.3, Tugtainer's OIDC authentication can still be initiated even when OIDC_ENABLED=false. The /auth/oidc/enabled endpoint correctly reports that OIDC is disabled. However, a direct request to /auth/oidc/login still starts the OIDC login flow, returns HTTP 302, sets an oidc_state cookie, and redirects the user to the configured OIDC authorization endpoint. This bypasses the intended OIDC disable switch. This issue has been patched in version 1.30.3.