CVE Vulnerability Catalog

Deserialization of untrusted data in Elasticsearch machine learning component can lead to remote code execution via object injection. A crafted trained model artifact could execute attacker-controlled logic with broader system-call surface. Requires authenticated user with privileges to create and deploy models.

Incorrect authorization in Kibana can lead to privilege escalation. A user with workflow edit permissions could cause scheduled workflow executions to run with higher-privileged user's privileges, allowing unauthorized data access and modification.

Konga before 2.1.0 has a privilege escalation vulnerability allowing low-privileged local attackers to execute arbitrary code by planting OpenSSL config or library files in a hardcoded filesystem path absent from default installations. On Windows, the missing directory is writable by any authenticated user, enabling attackers to create it and place malicious files that execute with the privileges of the launching account.

Wyoming before 1.10.2 has a server-side request forgery vulnerability allowing unauthenticated attackers with network access to force outbound connections to arbitrary targets by supplying a malicious `uri` query parameter to the HTTP API. Attackers can pass arbitrary `tcp://` or `unix://` URIs to endpoints including /api/info, /api/speech-to-text, and /api/text-to-speech to override the server-configured backend.

gRPC-Go prior to 1.83.1 stores each fragmented HTTP/2 DATA frame as a separate recvMsg, so millions of one-byte frames can consume disproportionate heap memory. An unauthenticated remote attacker can use concurrent multiplexed streams to exhaust process memory and cause a runtime panic or OOM termination. Receive-buffer compaction is enabled by default and can be controlled with GRPC_GO_EXPERIMENTAL_ENABLE_RECEIVE_BUFFER_COMPACTION. Fixed in 1.83.1.

Cleartext storage of sensitive information in the @step and @remote decorator pipeline component in Amazon SageMaker Python SDK before v3.11.0 and v2.256.0 might allow an authenticated remote user to extract the HMAC signing key from DescribePipeline API responses and forge valid integrity signatures for crafted function payloads, achieving code execution in another user's pipeline execution context within the same AWS account.

llama.cpp through commit 97f06e9, when started with the --reranking flag, allows remote attackers to cause a denial of service (std::bad_alloc and HTTP 500) via a negative top_n value in a POST request to /rerank.

llama.cpp b5693 and before has a Reachable Assertion via the gguf_reader::read function.

llama.cpp b5693 and before is vulnerable to Uncontrolled Recursion in common/json-schema-to-grammar.cpp, resulting in a denial of service.

An issue in kamailio v.6.1.1 and before allows a remote attacker to cause a denial of service via the ims_registrar_pcscf module, specifically the pcscf_save_pending/save_pending path and security-agreement parsing in sec_agree.c:parse_sec_agree().

A vulnerability in Kamailio version 6.1.1 and earlier allows a remote attacker to cause a denial of service via the IMS P-CSCF registration handling components.

An eval() injection vulnerability in the get_list function in modules/meta_parser.py in Fooocus 2.1.854 through 2.5.5 allows remote attackers to execute arbitrary Python code via a crafted styles payload in the EXIF metadata of an uploaded image file.

OpenAI Codex CLI for Windows, macOS, and Linux and Codex Desktop for Windows and macOS automatically collected Git repository metadata without disabling the repository-local core.fsmonitor setting. If a user opens or uses an attacker-prepared repository whose preserved .git/config sets core.fsmonitor to an attacker-controlled filesystem-monitor helper, Git can execute that helper while Codex collects repository metadata. The helper runs outside Codex's command sandbox and without a user-approval prompt, allowing attacker-controlled code to run with the user's privileges. The code can read, change, or delete the user's files and access other resources available to the user's account. An ordinary Git clone does not preserve the source repository's local .git/config; exploitation requires a repository delivered or copied with that configuration intact.

OpenAI Codex CLI for Windows, macOS, and Linux and Codex Desktop for Windows and macOS misclassified certain PowerShell commands as safe because their command-safety parser interpreted PowerShell's stop-parsing token (--%) differently than PowerShell itself. If a user opens an attacker-prepared repository and Codex follows its instructions, Codex can run a file-writing Git command without requesting user approval. On macOS and Linux, exploitation additionally requires separately installed PowerShell Core (pwsh) to be invoked. If filesystem protections permit the write, the command can modify Codex's configuration. If Codex later loads the modified configuration, it can launch an attacker-controlled MCP server and execute code with the user's privileges, allowing it to read, change, or delete files accessible to that account. The approval bypass does not disable filesystem sandboxing; the default filesystem sandbox on macOS and Linux can prevent writes outside permitted locations.

OpenAI Codex Desktop for Windows and macOS could execute attacker-controlled Git hooks because automated Git operations trusted the repository's local core.hooksPath setting. If a user opens an attacker-prepared repository whose preserved .git/config points core.hooksPath to an attacker-controlled directory, Codex can run a malicious hook while processing the repository. The hook executes outside Codex's command sandbox, without user approval, and with the user's privileges, allowing it to read, change, or delete the user's files and access other resources available to the user's account. An ordinary Git clone does not preserve the attacker-controlled repository-local configuration required for exploitation.

A vulnerability in the affected products allows an attacker to create a project and become its administrator. If exploited, the attacker could create, modify, and delete their own projects.

A data exposure vulnerability exists in the affected product. Hardcoded links in the source code lead to JSON files that can be accessed without authentication. If exploited, an attacker could view customer data.

Dell PowerStore contains an incorrect authorization vulnerability. A low privileged attacker with remote access could potentially exploit this vulnerability, leading to elevation of privileges.

A Broken Object Level Authorization vulnerability exists in Grashjs Atlas CMMS prior to v1.6.0. An authenticated user from one tenant can read and modify another tenant's company record by changing only the numeric ID in the /company/{id} endpoint. The application does not enforce tenant-level ownership checks when accessing or updating company objects, allowing cross-tenant access and modification of company profile data.

An issue in cleverange_auth version 0.1.10 allows a remote attacker to cause a denial of service via the account_verification function and the accounts/models.py component.