| CVE |
Vendors |
Products |
Updated |
CVSS v3.1 |
| llama.cpp builds b4882 through b9058 contain a heap buffer overflow vulnerability in the KV cache state restore path where the state_read_data() function computes write size without overflow checking, allowing attackers with write access to the slot_save_path directory to corrupt heap memory. Attackers can craft malicious state files where cell_count multiplication overflows or exceeds tensor buffer allocation to write attacker-controlled bytes past buffer boundaries, potentially resulting in heap metadata corruption, model weight corruption, or arbitrary code execution via function pointer overwrite. |
| llama.cpp builds b7492 through the latest b9060 contains a use-after-free vulnerability in the vocab pointer of llama-server when the --sleep-idle-seconds feature is enabled, allowing unauthenticated remote attackers to execute arbitrary code. Attackers can trigger the vulnerability by sending requests to affected endpoints while the server transitions to sleep mode, causing concurrent worker threads to dereference a freed vocab pointer that can be reclaimed with attacker-controlled data to achieve remote code execution. |
| llama.cpp builds b7492 through the latest b9060 contains a use-after-free vulnerability in llama-server affecting six tokenization endpoints (/tokenize, /detokenize, /infill, /apply-template, /rerank, and /anthropic/count_tokens) that bypass the task queue and access ctx_server.vocab directly on HTTP worker threads. Attackers can exploit a time-of-check-time-of-use race condition where the main thread destroys and frees vocab after the synchronization lock is released but before the handler finishes using it, causing a crash or potential code execution when --sleep-idle-seconds is configured. |
| Contiki-NG's MQTT client parse_publish_vhdr() in os/net/app-layer/mqtt/mqtt.c sets topic_len_received=1 before checking topic_len against the 64-byte limit, so an over-length topic returns early but leaves the flag set. On the next TCP segment, tcp_input() re-invokes the parser with topic_received==0, and the persisted topic_len_received==1 skips the length-reading block containing the guard, falling through directly to a memcpy() that uses the unvalidated 16-bit topic_len as the copy length. The 65-byte topic[] destination overruns into adjacent struct fields including the payload_chunk pointer, which subsequent MQTT code dereferences, giving a compromised or attacker-controlled broker an arbitrary-pointer-write primitive. Contiki-NG's MQTT implementation has no TLS support so the connection is plaintext. Impact ranges from information disclosure and denial of service to remote code execution on embedded targets without memory protection. |
| Dinky's POST /download/uploadFromRsByLocal handler passes the caller-supplied path parameter directly to new File(path) and file.transferTo(dest) with no path validation. The route is marked @SaIgnore and /download/** is excluded from the Sa-Token interceptor, so the only guard is a header equality check against a dinkyToken value whose default (efda1551-7958-4e0f-80a8-dfd107df3e38) is hardcoded in source and shipped to every deployment. Anyone who can reach Dinky's HTTP port (8888 by default) and supplies the hardcoded token can write arbitrary files as the Dinky service account. The default Docker image runs on 8888 with no proxy or authentication and chmod 777 on /opt/dinky, so the application's own classpath, launch scripts, and static assets are writable. Demonstrated impact: overwriting /opt/dinky/config/static/index.html served attacker JavaScript to admin browsers immediately, and writing /opt/dinky/org/dinky/Dinky.class executed attacker code as the Dinky service account at the next JVM start via a classpath-shadow launched by script/bin/auto.sh. Writes are uid 9999 (flink), not root, so /etc, /root, /home, and /usr are refused. Affects Dinky v1.2.5 (the current release) and the development branch, where the code is byte-identical. |
| FrontMCP is a TypeScript-first framework for the Model Context Protocol (MCP). Prior to 1.5.7, the sandboxed codecall:execute tool exposes live host Zod schema instances to the script via getTool(), and because Zod v4 defines _zod as a non-configurable, non-writable own property, the ECMAScript Proxy invariants force the security membrane to hand back the raw host object, letting a script reach _zod.constr.constructor (the host Function constructor) and execute arbitrary code in the server process. A single tools/call is sufficient to escape the sandbox and achieve remote code execution as the server user, exposing everything the process holds such as OAuth client secrets, JWT_SECRET, session keys, database credentials, and cloud instance metadata. Because the framework's DEFAULT_AUTH_OPTIONS is public mode, an unconfigured server serves this to unauthenticated callers, and on authenticated servers an indirect prompt injection in tool output or fetched content can trigger it without a human attackerThis issue is fixed in version 1.5.7. |
| PyAthena prior to 3.35.4 contains a sql injection vulnerability that allows unauthenticated attackers to inject arbitrary SQL by exploiting improper quote-escaping in DefaultParameterFormatter.format(), which routes DELETE and CTAS statements to the _escape_hive function that backslash-escapes single quotes rather than doubling them. Because Athena and Trino do not treat backslashes as escape characters inside string literals, attacker-supplied input such as a single quote followed by SQL syntax causes the parser to terminate the string literal prematurely, enabling data exfiltration via UNION SELECT, execution of destructive statements, and attacker-controlled CTAS destination and content. |
| Ground Station prior to 0.6.0 contains an unauthenticated blind server-side request forgery vulnerability in the orbital-source configuration path that allows any unauthenticated Socket.IO client to cause the ground-station process to issue outbound HTTP requests to attacker-chosen destinations. Attackers can connect to the Socket.IO server on port 7000 without credentials due to disabled authentication enforcement and a wildcard CORS policy, then submit a data_submission event with submit-orbital-sources action to persist an attacker-supplied URL in the database, then trigger an orbital sync via the equally unauthenticated background_task:start event. The URL is stored with no scheme allowlist, no host validation, and no rejection of loopback, RFC1918, or link-local (cloud instance metadata at 169.254.169.254) addresses, and is passed directly to requests.get in _fetch_http_3le and _fetch_http_omm in backend/tlesync/source_adapters.py. HTTP status codes and error messages from the outbound request are emitted in the orbital_sync_state Socket.IO event to all connected clients, providing a serviceable oracle for interpreting internal-service and cloud-metadata responses even though the raw response body is not directly leaked. Because the malicious source persists in the database across restarts and re-fires every 24 hours on the scheduled sync cycle, the primitive gives durable long-term SSRF without the attacker needing to remain connected. |
| OpenChamber 1.11.7 contains a path traversal vulnerability in the file-serving endpoints /api/fs/read, /api/fs/stat, and /api/fs/raw that allows unauthenticated remote attackers to read arbitrary files by supplying the allowOutsideWorkspace=true query parameter alongside an absolute path, bypassing the workspace boundary check in resolveReadPathFromContext. Attackers can exploit the vacuous isPathWithinRoot guard to read sensitive files such as the JWT signing secret, SSH private keys, API credentials, and environment variables, enabling full authentication bypass by forging session cookies on password-protected deployments. |
| OpenChamber 1.11.7 contains an unauthenticated remote code execution vulnerability that allows remote attackers to execute arbitrary shell commands by sending crafted POST requests to the /api/fs/exec endpoint, which passes commands verbatim to Node.js spawn() without any allowlist, blocklist, or argument validation. The authentication middleware becomes a no-op when UI_PASSWORD is not configured, matching the default Docker deployment, enabling attackers to execute arbitrary OS commands as the application user and retrieve full command output including stdout, stderr, and exit code from the server response. |
| Flarum before 1.8.16 contains a password reset token expiry bypass vulnerability that allows unauthenticated attackers to reuse expired password reset tokens by submitting them directly to the reset processing endpoint. The SavePasswordController::handle() method calls PasswordToken::findOrFail() without performing any expiry validation, allowing attackers to bypass the 24-hour token lifetime enforced only during form rendering and change any account's password to gain an authenticated session. |
| Flowise is a drag & drop user interface to build a customized large language model flow. Prior to 3.1.3, the POST /api/v1/oauth2-credential/refresh/:credentialId endpoint is included in WHITELIST_URLS and requires no authentication. The endpoint decrypts the stored credential, sends a refresh request to the configured OAuth provider with the client secret and refresh token, and returns the refreshed access_token in the response body. An attacker with a credential ID can use the token to access the victim's connected service and can also exhaust refresh-token quota. This issue is fixed in 3.1.3. |
| MaxSite CMS contains a remote code execution vulnerability that allows unauthenticated attackers to inject arbitrary PHP code into the application configuration file by submitting crafted POST requests to the install endpoint after installation is complete. Attackers can supply a malicious db_dbprefix value containing a single quote to break out of a PHP string literal in application/config/database.php, appending attacker-controlled PHP statements that are executed by the web server on every subsequent request, resulting in persistent unauthenticated remote code execution as the web-server process user. |
| Improper Privilege Management in Google SecOps (Chronicle SOAR) versions prior to 6.3.85 on Google Cloud Platform allows an authenticated attacker to escalate privileges to system-level administrative access using a crafted internal authentication header.
This vulnerability was patched with version 6.3.85, and no customer action is needed. |
| kotaemon through 0.12.0 contains an insecure deserialization vulnerability in the check_connection endpoint that allows unauthenticated attackers to instantiate arbitrary Python classes by supplying crafted YAML/JSON input with a __type__ field. Attackers can exploit this to override the __type__ field with subprocess.check_output and arbitrary arguments, achieving remote code execution with application process privileges. |
| Flowise is a drag-and-drop user interface for building customized large language model (LLM) flows. Prior to version 3.1.3, several custom-tool components — AgentAsTool, ChatflowTool, and ExecuteFlow — ran code in the in-process vm2 sandbox. To build that code, they inserted a user-controlled baseURL value straight into the JavaScript source, for example const url = "${baseURL}/..."; . The only check on baseURL was isValidURL , but a valid-looking URL can still contain characters that break out of a code string. An authenticated user could craft a baseURL that passed this check, closed the surrounding string, and injected their own JavaScript into the sandboxed script (code injection, CWE-94). The vm2 sandbox runs in the same Node.js process as Flowise and exposes risky dependencies. As a result, the injected code could escape the sandbox and run arbitrary code on the Flowise server as the Flowise process user. Exploitation only requires an authenticated session. The issue is fixed in version 3.1.3, which passes the URL to the sandbox as data instead of inserting it into code and adds stricter URL validation. |
| MaxSite CMS contains a PHP object injection vulnerability that allows unauthenticated attackers to execute arbitrary code by passing attacker-controlled serialized data in the maxsite_comuser cookie directly to unserialize() without validation or class allowlisting. Attackers can craft a malicious serialized PHP object payload delivered in a single HTTP request to trigger magic methods during object graph reconstruction, enabling property-oriented programming attacks or remote code execution via available gadget chains such as those targeting SoapClient or Imagick extensions. |
| Zbtlink router firmware ships an embedded remote-control implant, ENDLESSDOORS, present in every published build across the product line. It is the open-source ycsunjane/rctl tool built in as an OpenWrt package (librctl.so), started at boot and run as root under the process name kworker to blend in with the kernel's [kworker/*] threads. It opens no listening port; it phones home over cleartext TCP to a hardcoded command-and-control server (command channel 7000, interactive-shell callback 7001) with no authentication and no transport encryption, re-attempting contact roughly every 35 seconds. Its command handler passes any received string to popen() as uid=0, and a reserved rctlbash command returns an interactive root shell. Because the channel is unauthenticated and cleartext, control is not limited to whoever planted it: any party that answers at the C2 address, occupies the network path (DNS or route hijack), or acquires the hardcoded fallback domain obtains unauthenticated remote code execution as root. |
| Flowise is a drag & drop user interface to build a customized large language model flow. Prior to 3.1.3, a prompt injection sent to a chatflow using a CSV Agent node can cause the LLM to respond with a malicious Python script that bypasses the blocklist validator and executes in an unsandboxed Pyodide environment. The specific flaw exists within the run method of the CSV_Agents class, where untrusted data is used to construct an LLM prompt and the resulting pythonCode is validated by validatePythonCodeForDataFrame before execution. An attacker can leverage this to execute arbitrary code in the context of the service account. This issue is fixed in 3.1.3. |
| MaxSite CMS 109.5 and earlier contains an authentication bypass vulnerability in the AJAX dispatcher that allows unauthenticated attackers to access admin-gated endpoints by supplying any X-Requested-With header and requesting a base64-encoded path resolving to any *-ajax.php file in the codebase. Attackers can exploit this dispatcher bypass to reach privileged plugin endpoints without credentials, enabling actions such as manipulating poll states and vote counts, and amplifying the impact of any dangerous operation performed by admin-only ajax files across the plugin tree. |
