CVE Database

NVIDIA DGX Spark contains a vulnerability in the system firmware, where a privileged attacker could be able to cause an out-of-bounds write. A successful exploit of this vulnerability may lead to code execution, escalation of privileges, denial of service, information disclosure, and data tampering.

NVIDIA DGX Spark contains a vulnerability in the system firmware, where a privileged attacker could be able to cause a NULL pointer dereference. A successful exploit of this vulnerability may lead to code execution, escalation of privileges, denial of service, information disclosure, and data tampering.

NVIDIA DGX Spark contains a vulnerability in the system firmware, where a privileged attacker could be able to cause an out-of-bounds write. A successful exploit of this vulnerability may lead to code execution, escalation of privileges, denial of service, information disclosure, and data tampering.

NVIDIA UFM Enterprise contains a vulnerability in the web interface authorization component, where an authenticated user could cause improper authentication by sending specially crafted HTTP requests. A successful exploit of this vulnerability might lead to code execution and escalation of privileges.

NVIDIA UFM Enterprise contains a vulnerability in the plugin management API, where an authenticated user with low privileges could inject code by sending a specially crafted API request. A successful exploit of this vulnerability might lead to code execution, escalation of privileges and information disclosure.

The Kaltura HTML5 player (mwEmbed / html5lib) contains a local file disclosure vulnerability due to improper validation of the ServiceUrl parameter in mwEmbedLoader.php. This parameter is used as the base URL for a backend request and accepts non‑HTTP schemes such as file://. When an exception or error occurs, the response is subsequently deserialized and its raw contents are reflected to the client in an error message; this enables an unauthenticated, remote attacker to read any arbitrary internal file reachable by the server. Affected versions include html5lib v2.45, v2.103 and earlier, and other v2.x releases exposing the vulnerable endpoint.

There is a memory corruption vulnerability recently discovered in NI LabVIEW that may result in information disclosure or arbitrary code execution.  Successful exploitation requires an attacker to get a user to open a specially crafted VI.  This vulnerability affects NI LabVIEW 2026 Q3 (26.3.0) and prior versions.

There is a memory corruption vulnerability recently discovered in NI LabVIEW that may result in information disclosure or arbitrary code execution.  Successful exploitation requires an attacker to get a user to open a specially crafted VI.  This vulnerability affects NI LabVIEW 2026 Q3 (26.3.0) and prior versions.

Vocos instantiates a class named by a configuration file without restricting which class may be named. instantiate_class in vocos/pretrained.py takes the class_path value from the configuration, splits it into a module and an attribute, imports the module with __import__, resolves the attribute with getattr, and calls the result as args_class(*args, **kwargs) where kwargs is the config's own init_args mapping. No allowlist constrains the dotted path, so a configuration may name any importable callable and supply the arguments it is called with. Vocos.from_hparams reaches this for each of the feature_extractor, backbone and head entries, and Vocos.from_pretrained reaches it with a remote file: it downloads config.yaml from a caller-named Hugging Face repository and passes it straight to from_hparams. Loading a model from a repository the user does not control therefore executes code of the repository owner's choosing in the loading process. The neighbouring torch.load of the downloaded weights is a separate matter and is constrained on PyTorch releases that default weights_only to true, which leaves this path as the reachable one.

Winter CMS versions before 1.2.13 contain an incomplete fix for a Twig sandbox escape vulnerability in System\\Twig\\SecurityPolicy that allows authenticated backend users with template-editing permissions to bypass sandbox restrictions. Attackers can exploit method forwarding through Eloquent models and query builders using methods like saveQuietly(), deleteQuietly(), increment(), decrement(), and newQuery() to read and modify arbitrary database records, execute arbitrary SQL, and achieve remote code execution by injecting PHP into template code sections.

Nokogiri versions before 1.19.3 contain regular expression denial of service vulnerabilities in the CSS selector tokenizer affecting string-literal and identifier tokenization. Attackers can inject adversarial CSS selectors into methods like Node#css, Node#at_css, and Searchable#search to cause exponential regex backtracking and denial of service.

NLTK versions before 3.10.3 contain a path sandbox bypass vulnerability in corpus-reader constructors that allows attackers to read files outside the intended data root. Attackers can supply arbitrary corpus root paths to LinThesaurusCorpusReader and PanLexLiteCorpusReader constructors to access filesystem content and SQLite databases outside the pathsec sandbox boundary.

mcp-shell is an MCP server for running shell commands securely, auditably, and on demand. Prior to 0.6.0, the default security.yaml allows /usr/bin/git, while security.go omits ! from containsShellMetacharacters and containsDangerousShellConstructs and applies no per-executable argument policy. A caller of the shell_exec MCP tool can provide the command argument /usr/bin/git -c alias.pwn=!<arbitrary-command>, causing Git to create a shell alias and execute arbitrary OS commands as the mcp-shell process user. The default Docker image runs as mcpuser with Git installed and secure mode enabled, so the bypass is exploitable in the default deployment without additional authentication beyond MCP connectivity. This issue is fixed in version 0.6.0.

mcp-shell is an MCP server for running shell commands securely, auditably, and on demand. Prior to 0.6.0, the default Docker security.yaml includes /bin/bash in allowed_executables, while security.go validates only the first token and checkBlockedPatternsAndCommands does not reject the shell command-mode flag -c. A caller of the shell_exec MCP tool can provide the command argument `/bin/bash -c <arbitrary-command>`, which passes validation and reaches executor.go, where parseCommand and exec.CommandContext execute the arbitrary command as mcpuser outside the intended allowlist. This issue is fixed in version 0.6.0.

mcp-shell is an MCP server for running shell commands securely, auditably, and on demand. Prior to 0.6.0, config.go initializes Security.Enabled to false, and when MCP_SHELL_SEC_CONFIG_FILE is unset, main.go starts the documented bare-binary deployment without a security policy. SecurityValidator.validateCommand in security.go then short-circuits and allows every command supplied to the shell_exec MCP tool, so an LLM connected over stdio can execute unrestricted OS commands as the mcp-shell process user. The README from-source installation and MCP client configuration omit MCP_SHELL_SEC_CONFIG_FILE, making the insecure state the documented default. This issue is fixed in version 0.6.0.

PraisonAI is a multi-agent teams system. Prior to praisonai 4.6.51, the Jobs API create_app function mounts /api/v1/runs without authentication. Any reachable caller can submit jobs, read results, cancel runs, or delete jobs using operator credentials. The fix adds PRAISONAI_JOBS_API_KEY middleware for Authorization or X-API-Key. This issue is fixed in version 4.6.58.

PraisonAI is a multi-agent teams system. Prior to praisonai 4.6.58, create_auth_middleware() allows requests when auth=api-key lacks PRAISONAI_API_KEY or JWT authentication lacks PRAISONAI_JWT_SECRET. An externally bound Recipe server can therefore accept unauthenticated POST /v1/recipes/run requests despite authentication being enabled. This issue is fixed in version 4.6.58.

PraisonAI is a multi-agent teams system. Prior to praisonai 4.6.58, MCP HTTP Stream _validate_origin uses request_origin.startswith(allowed), allowing the attacker-controlled localhost.attacker.com HTTP origin to satisfy the localhost allowlist. A webpage can send Content-Type: text/plain requests without preflight and invoke tools/call without an API key, including file writes that persist agent instructions. This issue is fixed in version 4.6.58.

A user who can read an existing remote VCS repository can replace its configured origin or supply an absolute VCS data URL.

An authenticated user may initiate repository migration operations without required repository permissions, potentially causing information disclosure, unauthorized state changes, and service disruption. Fixed versions address the issue.

PraisonAI is a multi-agent teams system. Prior to praisonai 4.6.58, praisonai serve agents and praisonai serve unified parse --api-key but _create_agents_app() and _create_unified_app() do not install a credential check. Unauthenticated callers can reach POST /agents and POST /api/v1/agents/{id}/invoke. This issue is fixed in version 4.6.58.

PraisonAI is a multi-agent teams system. Prior to praisonai 4.6.51, is_path_within_directory() uses os.path.abspath() rather than os.path.realpath() for the workspace boundary. A symlink inside workspace can point outside and still pass the check, allowing read_file and other code tools to access files outside the configured workspace. This issue is fixed in version 4.6.58.

PraisonAI is a multi-agent teams system. Prior to praisonai 4.6.51, praisonai serve agents parses config["api_key"] but _create_agents_app() does not authenticate POST /agents or POST /agents/{agent_name}. Missing or incorrect bearer and X-API-Key values still reach agent execution. This issue is fixed in version 4.6.58.

PraisonAI is a multi-agent teams system. Prior to praisonai 4.6.58, JobSubmitRequest.validate_webhook_url() accepts webhook_url when resolution raises socket.gaierror because the exception path uses except socket.gaierror: pass. JobExecutor._send_webhook() later performs a fresh lookup, allowing DNS changes to direct the request to an internal service. This issue is fixed in version 4.6.58.

PraisonAI is a multi-agent teams system. From praisonai 4.6.34 until 4.6.58, praisonai serve agents accepts --api-key but _create_agents_app() does not authenticate POST /agents or POST /agents/{agent_name}. A network caller can invoke configured agents without credentials even when an API key was supplied. This issue is fixed in version 4.6.58.

PraisonAI is a multi-agent teams system. Prior to praisonaiagents 1.6.58, AgentServer exposes ServerConfig.auth_token but AgentServer._create_app does not check it on any route. A remote caller can subscribe, publish, and perform other actions without a valid bearer token or X-Auth-Token even when authentication is configured. This issue is fixed in version 1.6.58.

PraisonAI is a multi-agent teams system. Prior to praisonaiagents 1.6.58, the FileMemory constructor joins unsanitized user_id into self.user_path. A caller supplying ../ or path separators can escape the memory directory and write JSON data to arbitrary process-writable locations. The fix sanitizes user_id before constructing self.user_path. This issue is fixed in version 1.6.58.

PraisonAI is a multi-agent teams system. Prior to praisonaiagents 1.6.58, spider_tools._host_is_blocked() does not resolve ordinary hostnames before scrape_page fetches them. A hostname such as 127.0.0.1.nip.io passes validation and resolves to loopback, permitting internal HTTP access. The fix uses socket.getaddrinfo and fails closed on DNS errors. This issue is fixed in version 1.6.58.

A flaw was found in sos clean, a utility within the sos package. This vulnerability allows a local attacker to perform arbitrary file creation or overwrite. By crafting a malicious tar archive, an attacker can exploit a path traversal issue during tar extraction, where symlink and hardlink targets are not properly validated. This enables the attacker to write files to arbitrary locations on the system with the privileges of the sos clean process, which often runs as root.

A weakness has been identified in dekdee adobe-xd-mcp 1.0.0. Impacted is an unknown function of the file src/parsers/xd-parser.ts of the component file-access-from-request Endpoint. Executing a manipulation of the argument outputFile/outputDir can lead to path traversal. It is possible to launch the attack remotely. The exploit has been made available to the public and could be used for attacks. The project was informed of the problem early through an issue report but has not responded yet.