Dem00000

A vulnerability was found in tjake Jlama up to 0.8.4/0.9.0-beta. Affected by this vulnerability is the function SafeTensorIndex.loadWeights of the file jlama-core/src/main/java/com/github/tjake/jlama/safetensors/SafeTensorIndex.java. Performing a manipulation of the argument weight map results in path traversal. The attack may be initiated remotely. The exploit has been made public and could be used. The project was informed of the problem early through an issue report but has not responded yet.

A vulnerability has been found in braindecode up to 1.8.1. Affected is the function resolve type kwargs of the file braindecode/models/util.py of the component Model Config Deserialization. Such manipulation of the argument Activation leads to use of externally-controlled input to select classes or code. The attack can be launched remotely. The exploit has been disclosed to the public and may be used. Upgrading to version 2.0.0 is able to address this issue. The name of the patch is 05b764d93b9a50400f523206fbe03d73d0cf5edb. The affected component should be upgraded.

A weakness has been identified in Turboderp-org ExLlamaV3 up to 1.6.3. The impacted element is the function routing sel norm kernel of the file exllamav3/exllamav3 ext/routing.cu of the component hash-MoE routing. This manipulation of the argument gate.tid2eid causes out-of-bounds read. Remote exploitation of the attack is possible. Upgrading to version 1.6.4 is sufficient to resolve this issue. Patch name: c9a57ae832a3035e264e0a412e1daf4ebb699b98. Upgrading the affected component is recommended.

A security vulnerability has been detected in braindecode up to 1.8.2. This affects the function EEGModuleMixin. from pretrained of the file braindecode/models/base.py of the component Hub Integration. Such manipulation of the argument filename leads to path traversal. The attack can be executed remotely. The exploit has been disclosed publicly and may be used. Upgrading to version 2.0.0 mitigates this issue. The name of the patch is 05b764d93b9a50400f523206fbe03d73d0cf5edb. It is recommended to upgrade the affected component.

A security flaw has been discovered in huggingface finetrainers 0.1.0/0.2.0. The impacted element is the function resolve component cls of the file finetrainers/utils/model.py of the component Model Index Handler. Performing a manipulation of the argument Library results in code injection. The attack can be initiated remotely. The exploit has been released to the public and may be used for attacks. The project was informed of the problem early through an issue report but has not responded yet.

A weakness has been identified in lxtGH OMG-Seg up to f3b26b609bae919760cd3ccf09c36179bb8624a0. This affects the function load state dict into model of the file omg llava/xtuner/utils/handle moe load and save.py of the component MoE Shard Checkpoint Loader. Executing a manipulation of the argument weight map can lead to path traversal. The attack can be launched remotely. The exploit has been made available to the public and could be used for attacks. This product does not use versioning. This is why information about affected and unaffected releases are unavailable. The project was informed of the problem early through an issue report but has not responded yet.

A security flaw has been discovered in Turboderp-org ExLlamaV3 up to 1.6.3. The affected element is the function ngram dequant of the file exllamav3/exllamav3 ext/ngram.cu of the component NGramEmbedding. The manipulation of the argument head bias results in out-of-bounds read. The attack may be launched remotely. The exploit has been released to the public and may be used for attacks. Upgrading to version 1.6.4 is sufficient to fix this issue. The patch is identified as c9a57ae832a3035e264e0a412e1daf4ebb699b98. It is suggested to upgrade the affected component.

An issue exists in the Model Loading component within the `load pretrained model()` function located in the `videollama2/model/ init .py` file. Improper validation of the integrity check value occurs when the `name or path` argument is manipulated, allowing for remote exploitation.

Cross site scripting occurs due to the manipulation of the `template` argument within a function in the file argilla-frontend/components/features/annotation/container/fields/sandbox/Sandbox.vue. This allows for remote exploitation.

Remote cross-site scripting is possible through the manipulation of the `text` argument within the file argilla-frontend/components/features/annotation/container/fields/text-field/TextField.vue. Cross-site scripting is a technique where malicious scripts are injected into trusted websites.

A flaw in the Settings Page component allows remote attackers to perform cross-site scripting (XSS) by manipulating the `Name` argument within the `Dataset.from disk()` function located in the `argilla-frontend/components/features/annotation/settings/SettingsInfo.vue` file.

A weakness in the Pocket-TTS component allows an out-of-bounds write when the `mmproj` argument is manipulated. This issue occurs within the `list pockettts state slots()` function located in the tools/mtmd/models/pockettts-gen.cpp file. Exploitation requires local access.

An issue exists in the AutoBackend Model Loading component within the `eval()` function of the `ultralytics/nn/autobackend.py` file. Improper neutralization of directives in dynamically evaluated code occurs when the `imgsz`, `names`, or `kpt shape` arguments are manipulated. This flaw requires a local approach to exploit.

An issue exists in the AutoBackend Model Loading component within the `session.get modelmeta.custom metadata map` function of the `models/MobileSAMv2/ultralytics/nn/autobackend.py` file. Improper neutralization of directives in dynamically evaluated code occurs when the `imgsz`, `names`, or `kpt shape` arguments are manipulated. This flaw requires local access to be exploited.

An issue exists in the AutoBackend Model Loading component within the `session.get modelmeta.custom metadata map` function of the `ultralytics/nn/autobackend.py` file. Manipulation of the `imgsz`, `names`, or `kpt shape` arguments leads to improper neutralization of directives in dynamically evaluated code. Exploitation of this flaw requires local access.

An issue exists in the Model Loader component within the `whisper model load()` function of the `src/whisper.cpp` file. A local attacker can trigger an out-of-bounds read by manipulating the `ttype` argument. An out-of-bounds read occurs when the software reads data past the end of the intended buffer, potentially exposing sensitive information from memory.

An issue exists in the Model Loader component within the file external/whispercpp/whisper.cpp. Specifically, the `whisper model load()` function is susceptible to an out-of-bounds write, which occurs when the `n dims` argument is manipulated. This flaw allows for remote attacks.

An issue exists in the `retrieval lm` component within the `Indexer.deserialize from()` function of the `retrieval lm/src/index.py` file. A remote attacker can trigger deserialization by manipulating the `index meta.faiss` argument. Deserialization is a process of converting a serialized format (like a byte stream) back into an object, which can lead to arbitrary code execution if the input is untrusted.

An OS command injection issue exists within the Android WebView Bridge component, specifically in an unknown function located in the `android/app/src/main/java/com/openclaw/android/JsBridge.kt` file. This flaw allows for the execution of arbitrary operating system commands, although the attack can only be performed locally.

An issue exists in the Cache Handler component within the `request a()` function of the `ad miner/scripts/analyse cache.py` file. Local manipulation of the `sys.argv[1]` variable leads to insecure deserialization, which occurs when data is converted from a byte stream back into an object. This flaw can only be exploited by a local attacker.