28 matches found
Stored XSS in Prompt-Sharing Direct Messages leading to account takeover
This report is not public...
Uncontrolled Recursion in NLTK StupidBackoff Language Model Allows Denial of Service
Description NLTK’s StupidBackoff language model contains an unbounded recursive call in its scoring logic. The recursion depth is directly proportional to the length of the supplied context and is not limited or validated by the library. This behavior is triggered through normal, documented API...
Integer Overflow lead to DOS in handling Accept-Encoding header in API /v2/models/<model-name>/generate
This report is not public...
Arbitrary Code Execution in NLTK StanfordSegmenter via untrusted JAR loading
Description nltk.tokenize.StanfordSegmenter dynamically loads external Java .jar files without verification or sandboxing. If an attacker can supply or replace the JAR e.g., a poisoned model download, MITM package swap, or dependency poisoning, arbitrary Java bytecode executes at import time...
Zip Slip Vulnerability in NLTK Downloader Leading to Remote Code Execution
The NLTK downloader component contains a critical vulnerability where the unzipiter function uses zipfile.extractall without path validation or security checks. This allows malicious zip packages to achieve arbitrary code execution when extracted and loaded by NLTK. The vulnerability exists becau...
Dependacy chain attack through hijacking broken github repository at https://github.com/huggingface/transformers/blob/main/src/\ntransformers/models/fuyu/\nconvert_fuyu_model_weights_to_hf.py
Description Type: Dependency Chain Attack through hijacking broken github repository Risk: High Allows arbitrary code execution in model conversion workflows Affected Asset: https://github.com/adept-ai-labs/adept-inference Broken URL in Hugging Face Transformers Root Cause The Hugging Face...
Brotli decompression bomb DoS
Description urllib3 can not stream brotli-encoded responses properly unlike the way it handles gzip responses. It always loads entire decompressed response body into memory when reading brotli-encoded response, which allows malicious servers to perform DoS attack by responding with decompression...
Full system file read and delete via GET /api/v1/images/download/{bulk_download_item_name}
Description For invokeai version v6.0.0a1 and below, there is an endpoint for bulk downloading zip file. With some manipulation of the filename arguments, attacker can read and also delete any files on the server through this endpoint. P/S: Tested on Windows Proof of Concept Request: GET...
Partial Account Takeover due to Insecure Data Querying
This report is not public...
Arbitrary file deletion on Windows via the '/v1/agent/hub/update' endpoint.
This report is not public...
Integer Overflow In /v2/repository/models/<model_name>/load
This report is not public...
malicious gguf model can be uploaded and created causing division by zero via network, leading to DoS
This report is not public...
DoS using malicious gguf model file
This report is not public...
Denial of service through tracking and requesting Aim objects through web API
This report is not public...
dify tools vanna has pandas query inject
This report is not public...
Denial of service by tracking large images
This report is not public...
Ollama server is vulnerable to OOM DoS attacks when using `makeRequestWithRetry` and `getAuthorizationToken` functions
This report is not public...
(Blind) Stored XSS through the debug_log.html generated by the Latex Proof-Reading Module
This report is not public...