Writeup Exploits
62,541 exploits tracked across all sources.
Guardrails AI Hub <0.6.7 - Code Injection
Guardrails AI thru 0.6.7 contains a code injection vulnerability (CWE-94) in its Hub package installation mechanism. When installing validator packages via guardrails hub install, the system retrieves a manifest from the Guardrails Hub and dynamically executes a script specified in the post_install field. The script path is constructed from untrusted manifest data and executed without proper validation or sanitization, allowing remote code execution. An attacker who can publish malicious packages to the Hub can inject arbitrary code that will be executed on any system where a victim installs the malicious package.
CVSS 9.8
Horovod <= 0.28.1 - Unauthenticated Remote Code Execution via Insecure KVStore Deserialization
Horovod thru 0.28.1 contains an insecure deserialization vulnerability (CWE-502) in its KVStore HTTP server component. The KVStore server, used for distributed task coordination, lacks authentication and authorization controls, allowing any remote attacker to write arbitrary data via HTTP PUT requests. When a Horovod worker reads data from the KVStore (via HTTP GET), it deserializes the data using cloudpickle.loads() without verifying its source or integrity. An attacker can exploit this by sending a malicious pickle payload to the server before the legitimate data is written, causing the victim worker to deserialize and execute arbitrary code, leading to remote code execution.
CVSS 9.8
imgaug <= 0.4.0 - Remote Code Execution via Insecure Pickle Deserialization in BackgroundAugmenter
The imgaug library thru 0.4.0 contains an insecure deserialization vulnerability in its BackgroundAugmenter class within the multicore.py module. The class uses Python's pickle module to deserialize data received via a multiprocessing queue in the _augment_images_worker() method without any safety checks. An attacker who can influence the data placed into this queue (e.g., through social engineering, malicious input scripts, or a compromised shared queue) can provide a malicious pickle payload. When deserialized, this payload can execute arbitrary code in the context of the worker process, leading to remote or local code execution depending on the deployment scenario.
CVSS 9.8
llm CLI < 0.27.1 - Remote Code Execution via --functions Argument
The llm CLI tool thru 0.27.1 contains a critical code injection vulnerability via its --functions command-line argument. This argument is intended to allow users to provide custom Python function definitions. However, the tool directly executes the provided code using the unsafe exec() function without any sanitization, sandboxing, or security restrictions. An attacker can exploit this by crafting a malicious llm command with arbitrary Python code in the --functions argument and using social engineering to trick a victim into running it. This leads to arbitrary code execution on the victim's system, potentially granting the attacker full control.
CVSS 9.8
Ludwig Framework <=0.10.4 - Deserialization
The Ludwig framework thru 0.10.4 is vulnerable to insecure deserialization (CWE-502) in its model serving component. When starting a model server with the ludwig serve command, the framework loads model weight files using torch.load() without enabling the security-restrictive weights_only=True parameter. This default behavior allows the deserialization of arbitrary Python objects via the pickle module. An attacker can exploit this by providing a maliciously crafted PyTorch model file, leading to arbitrary code execution on the system hosting the Ludwig model server.
CVSS 9.8
Ludwig Framework <=0.10.4 - Deserialization
The Ludwig framework thru 0.10.4 is vulnerable to insecure deserialization (CWE-502) through its predict() method. When a user provides a dataset file path to the predict() method, the framework automatically determines the file format. If the file is a pickle (.pkl) file, it is loaded using pandas.read_pickle() without any validation or security restrictions. This allows the deserialization of arbitrary Python objects via the unsafe pickle module. A remote attacker can exploit this by providing a maliciously crafted pickle file, leading to arbitrary code execution on the system running the Ludwig prediction.
CVSS 9.8
mamba < 2.2.6 - Remote Code Execution via Insecure Model Deserialization
The mamba language model framework thru 2.2.6 is vulnerable to insecure deserialization (CWE-502) when loading pre-trained models from HuggingFace Hub. The MambaLMHeadModel.from_pretrained() method uses torch.load() to load the pytorch_model.bin weight file without enabling the security-restrictive weights_only=True parameter. This allows the deserialization of arbitrary Python objects via the pickle module. An attacker can exploit this by publishing a malicious model repository on HuggingFace Hub. When a victim loads a model from this repository, arbitrary code is executed on the victim's system in the context of the mamba process.
CVSS 9.8
mem0 v1.0.0 - Unauthenticated SQL Injection via DELETE /memories Endpoint
The mem0 v1.0.0 server lacks authentication and authorization controls for its memory reset functionality accessible via the DELETE /memories endpoint. An unauthenticated attacker can send a DELETE request that triggers a reset operation, leading to the execution of a DROP TABLE SQL statement. This results in the deletion of the entire memory database table, causing catastrophic data loss and a complete denial of service for all users of the service.
CVSS 9.1
mem0 1.0.0 - Unauthenticated Memory Deletion via DELETE /memories Endpoint
The mem0 1.0.0 server lacks authentication and authorization controls for its memory deletion API endpoint (DELETE /memories). The endpoint allows unauthenticated users to delete memory records by specifying arbitrary user identifiers (e.g., user_id, run_id, agent_id) in the request query parameters. A remote attacker can exploit this by sending unauthenticated DELETE requests to erase memory data for any user, leading to unauthorized data loss and denial of service.
CVSS 6.5
mem0 1.0.0 - Unauthenticated Memory Record Manipulation via Memory Management API
The mem0 1.0.0 server lacks authentication and authorization controls for its memory management API endpoints. Critical functions such as updating memory records (PUT /memories/{memory_id}) are exposed without any verification of the requester's identity or permissions. A remote attacker can exploit this by sending unauthenticated requests to modify, overwrite, or delete arbitrary memory records, leading to unauthorized data manipulation and potential data loss.
CVSS 7.5
mem0ai mem0 faiss.py pickle.dump deserialization
A vulnerability was found in mem0ai mem0 up to 1.0.11. This affects the function pickle.load/pickle.dump of the file mem0/vector_stores/faiss.py. Performing a manipulation results in deserialization. It is possible to initiate the attack remotely. The exploit has been made public and could be used. The patch is named 62dca096f9236010ca15fea9ba369ba740b86b7a. Applying a patch is the recommended action to fix this issue.
CVSS 6.3
mem0ai mem0 faiss.py pickle.dump deserialization
A vulnerability was found in mem0ai mem0 up to 1.0.11. This affects the function pickle.load/pickle.dump of the file mem0/vector_stores/faiss.py. Performing a manipulation results in deserialization. It is possible to initiate the attack remotely. The exploit has been made public and could be used. The patch is named 62dca096f9236010ca15fea9ba369ba740b86b7a. Applying a patch is the recommended action to fix this issue.
CVSS 6.3
mem0ai mem0 faiss.py pickle.dump deserialization
A vulnerability was found in mem0ai mem0 up to 1.0.11. This affects the function pickle.load/pickle.dump of the file mem0/vector_stores/faiss.py. Performing a manipulation results in deserialization. It is possible to initiate the attack remotely. The exploit has been made public and could be used. The patch is named 62dca096f9236010ca15fea9ba369ba740b86b7a. Applying a patch is the recommended action to fix this issue.
CVSS 6.3
SQL injection via channel name in Postgrex.Notifications.listen/3 and unlisten/3
Improper Neutralization of Special Elements used in an SQL Command ('SQL Injection') vulnerability in elixir-ecto postgrex ('Elixir.Postgrex.Notifications' module) allows SQL Injection.
The channel argument passed to 'Elixir.Postgrex.Notifications':listen/3 and 'Elixir.Postgrex.Notifications':unlisten/3 is interpolated directly into LISTEN "..." / UNLISTEN "..." SQL statements without escaping the " character. An attacker who can influence the channel name can inject a " to break out of the quoted identifier and append arbitrary SQL. Because the notifications connection uses the PostgreSQL simple query protocol, multi-statement payloads are accepted, allowing DDL and DML commands to be chained (e.g. ; DROP TABLE ...; --). The same unsanitized interpolation also occurs in handle_connect/1 when replaying LISTEN commands after a reconnect.
This vulnerability is associated with program file lib/postgrex/notifications.ex and program routines 'Elixir.Postgrex.Notifications':listen/3, 'Elixir.Postgrex.Notifications':unlisten/3, 'Elixir.Postgrex.Notifications':handle_connect/1.
This issue affects postgrex: from 0.16.0 before 0.22.2, from pkg:github/elixir-ecto/postgrex@266b530faf9bde094e31e0e4ab851f933fadc0f5 before 0.22.2.
CVSS 7.8
Horilla: Open Redirect via Unvalidated `next` Parameter in Notification Endpoints
Horilla is an HR and CRM software. In 1.5.0, the notification endpoints trust the unvalidated next parameter and redirect users to arbitrary external URLs. This allows an attacker to turn trusted application links into phishing or social-engineering redirects.
JunoClaw: upload_wasm accepted arbitrary filesystem paths without validation
JunoClaw is an agentic AI platform built on Juno Network. Prior to 0.x.y-security-1, the upload_wasm MCP tool accepted a filesystem path from the agent and uploaded whatever bytes the path resolved to, with no validation of location, symlink target, file size, or file format. This vulnerability is fixed in 0.x.y-security-1.
CVSS 8.5
JunoClaw: plugin-shell shell-metacharacter injection via shell wrapper
JunoClaw is an agentic AI platform built on Juno Network. Prior to 0.x.y-security-1, plugin-shell's run_command wrapped every agent-supplied command in 'sh -c' / 'cmd /C' and passed the full argument string to the shell's parser, allowing shell metacharacters in agent-supplied arguments to be interpreted as command syntax. This vulnerability is fixed in 0.x.y-security-1.
CVSS 8.4
JunoClaw: plugin-shell shell-injection bypass via substring blocklist
JunoClaw is an agentic AI platform built on Juno Network. Prior to 0.x.y-security-1, substring-based blocklist in plugin-shell's command-safety check could be bypassed by adversarial argument constructions, allowing unauthorized command execution on the host when combined with the companion advisory. Pre-patch, the check was applied to the raw command string rather than the parsed first token. This vulnerability is fixed in 0.x.y-security-1.
CVSS 8.4
JunoClaw: MCP write tools exposed raw BIP-39 mnemonic as a tool-call parameter
JunoClaw is an agentic AI platform built on Juno Network. Prior to 0.x.y-security-1, every MCP write tool (send_tokens, execute_contract, instantiate_contract, upload_wasm, ibc_transfer, etc.) accepted 'mnemonic: string' as an explicit tool-call parameter. The BIP-39 seed was consequently embedded in the LLM tool-call JSON, exposing it to any transport, log, or telemetry surface in the path between the LLM provider and the MCP process. This vulnerability is fixed in 0.x.y-security-1.
CVSS 9.8
JunoClaw: SSRF in WAVS computeDataVerify allows cloud-metadata and internal-service access
JunoClaw is an agentic AI platform built on Juno Network. Prior to 0.x.y-security-1, the WAVS bridge's computeDataVerify called fetch() on agent-supplied URLs without validating scheme, port, or resolved IP, resulting in an SSRF vulnerability. This vulnerability is fixed in 0.x.y-security-1.
CVSS 8.2
phpseclib: CVE-2024-27355 mitigation bypass — OID amplification DoS in ASN1::decodeOID()
phpseclib is a PHP secure communications library. Prior to 1.0.29, 2.0.54, and 3.0.52, anyone loading untrusted ASN1 files (eg. X509 certificates, RSA PKCS8 private or public keys, etc). This is a bypass of CVE-2024-27355. This vulnerability is fixed in 1.0.29, 2.0.54, and 3.0.52.
CVSS 7.5
WGDashboard < 4.3.2 - Unauthenticated Host File System Access
WGDashboard is a dashboard for WireGuard VPN. Prior to 4.3.2, there are critical vulnerabilities affecting WGDashboard that, if exploited, could allow unauthorized parties to access the host file system without authentication. This vulnerability is fixed in 4.3.2.
CVSS 9.8
Perl YAML::Syck <1.38 Base60 Parsing - Out-of-Bounds Read
YAML::Syck versions before 1.38 for Perl has an out-of-bounds read.
The base60 (sexagesimal) parsing code in perl_syck.h has a buffer underflow bug in both int#base60 and float#base60 handlers. When processing the leftmost segment of a colon-separated value (e.g., the 1 in 1:30:45), the inner while loop can decrement a pointer past the start of the string buffer:
while ( colon >= ptr && *colon != ':' )
{
colon--;
}
if ( *colon == ':' ) *colon = '\0'; // colon may be ptr-1 here
When no colon is found (final/leftmost segment), colon becomes ptr-1, and the subsequent *colon dereference reads one byte before the allocated buffer.
CVSS 7.3
LWP::UserAgent versions before 6.83 for Perl leak Authorization and Proxy-Authorization headers on cross-origin redirects
LWP::UserAgent versions before 6.83 for Perl leak Authorization and Proxy-Authorization headers on cross-origin redirects.
On a 3xx response, the redirect handler strips only Host and Cookie before issuing the follow-up request. Caller-supplied Authorization and Proxy-Authorization headers are sent unchanged to the redirect target, including across scheme, host, or port changes.
A redirect to an attacker controlled host therefore discloses the caller's credentials to that host.
CVSS 6.5
Grav: Administrative Account Disruption and Privilege De-escalation via User Overwrite Logic
Grav is a file-based Web platform. Prior to 2.0.0-beta.2, a business logic vulnerability in the Grav Admin Panel allows a low-privileged user (with only user creation permissions) to overwrite existing accounts, including the primary administrator. By creating a new user with a username that already exists, the system updates the existing account's metadata and permissions instead of rejecting the request. This leads to a Denial of Service (DoS) on administrative functions and Privilege De-escalation of the root account. This vulnerability is fixed in 2.0.0-beta.2.
CVSS 8.1
By Source