AWS VDP: Sandbox User Can Inject Rogue CA Certificate into OS Trust Store via Sudo-Allowed deploy-certificates.sh

A vulnerability was found in the Bedrock AgentCore Code Interpreter sandbox. The sandbox granted a user passwordless sudo access to a script that deployed certificates to the OS trust store. An attacker could have generated a rogue CA certificate, placed it in a writable directory, and then used...

AWS VDP: Incomplete Input Sanitization in CodeInterpreter install_packages Allows Command Injection via pip Flags

The Bedrock AgentCore Python SDK was found to have an incomplete input sanitization vulnerability in the installpackages method. The sanitization process failed to properly handle certain pip flags, such as --index-url, which allowed arbitrary command execution within the Code Interpreter sandbox...

AWS VDP: Bedrock AgentCore Starter Toolkit Creates Gateway IAM Roles Without Confused Deputy Protections

The Bedrock AgentCore Starter Toolkit was found to create IAM roles for the AgentCore Gateway with trust policies that lacked certain condition keys. This enabled a cross-account confused deputy attack where any AgentCore workload in any AWS account could assume the victim's Gateway role. The rol...

curl: Unbounded GZIP Decompression Leading to Event-Loop Starvation

When libcurl is configured to decompress HTTP responses via CURLOPTACCEPTENCODING or the --compressed CLI flag, it lacks decompression bounds checking or a mechanism to yield execution during massive expansion tasks. If an attacker provides a highly compressed payload zip bomb, libcurl's underlyi...

Arbitrary code execution via malicious logging configuration (dictConfig factory injection)

Kedro v1.2.0 passes user-controlled YAML config to logging.config.dictConfig without validation. The factory key enables arbitrary callable instantiation, achieving RCE at startup. kedro/framework/project/init.py, ProjectLogging: loggingconfig = Pathpath.readtextencoding="utf-8"...

AWS VDP: Kiro IDE Stores Auth Tokens with World-Readable Permissions (0644)

The Kiro IDE version 0.11.107 wrote authentication tokens access token and refresh token to a file with world-readable permissions 0644. The file contained sensitive information, including the access token, refresh token, and profile ARN. This exposed the credentials to potential unauthorized...

curl: HTTP/2 server push accepts a non-authoritative :scheme=https over cleartext h2c, enabling HTTPS cache-key poisoning

Summary: I found that libcurl 8.19.0 accepts an HTTP/2 pushed stream on a cleartext h2c connection even when the server sends :scheme=https in PUSHPROMISE. In lib/http2.c, settransferurl builds the pushed handle URL from the server-supplied :scheme, :authority, and :path, but PUSHPROMISE validati...

Shopify: admin.shopify.com: Shopify Flow continues sending internal emails to a configured recipient after the staff author is removed

Summary: A staff user removed from a Shopify store continues to receive customer Personally Identifiable Information PII via a previously created automation workflow. The workflow, configured using Shopify Flow, continues executing after the user's access has been revoked and sends sensitive...

curl: Security Vulnerability Report: Protocol Injection via Programmatic Options

Summary Multiple text-based protocol handlers in libcurl including FTP, SMTP, POP3, and IMAP are vulnerable to protocol command injection. This occurs when an application sets credentials or other protocol-specific options programmatically e.g., via CURLOPTUSERNAME, CURLOPTPASSWORD, or...

Node.js: Permission Model bypass via FileHandle.utimes() in the promises API

Vulnerability description not provided...

PortSwigger Web Security: Out of scope: Improper Input Validation Order on /api-internal/login via password field leads to unnecessary resource consumption

A security issue was discovered in the /api-internal/login authentication endpoint of the internal login interface of Burp Suite DAST Enterprise. The issue was caused by improper input validation order, where the application processed user-supplied input before enforcing field-level validation...

Path Traversal in DiskIOStore via Unsanitized Layer Names in Keras 3

Description A logic flaw in the Keras 3 v3.0.0+ model saving and loading library handles internal asset paths insecurely. Specifically, the DiskIOStore.make method in keras/src/saving/savinglib.py uses user-provided layer names to construct directory paths without sanitizing for parent directory...

Nextcloud: PIN bypass in PassCodeActivity via back button

A vulnerability was discovered in the PassCodeActivity of a certain application. The vulnerability allowed bypassing the PIN code by pressing the back button...

Pipelines Frontend /_proxy/ Unrestricted SSRF

This report is not public...

curl: HTTP/1.1 Response Desynchronization via conflicting CL/TE headers in Proxy CONNECT

Summary: curl fails to prioritize the Transfer-Encoding: chunked header over Content-Length in HTTP/1.1 proxy responses specifically 407/401 auth challenges, violating RFC 9112 Section 6.1. I have identified the root cause in cf-h1-proxy.c. In the response-handling loop around line 466, the code...

Cloudflare Public Bug Bounty: HTTP Request Smuggling via Unsanitized Hop-by-Hop Headers

The vulnerability allowed forwarding of HTTP/1 h2c upgrade requests to upstream servers, bypassing request-level filters applied by the proxy. The issue was addressed in Pingora version 0.9.0, which restricted HTTP/1 upgrades to WebSocket by default...

curl: CVE-2026-4873: connection reuse ignores TLS requirement

A vulnerability was discovered in libcurl's connection reuse for cleartext-upgrade mail protocols. The vulnerability was that the later transfer's CURLOPTUSESSL option was not properly included if a plaintext connection was already open and reusable. This affected the smtp://, pop3://, and imap:/...

Monero: ZMQ RPC Log Injection and Untrusted Payload Persistence

A vulnerability was reported in the Monero CLI daemon where the ZMQ RPC request path logs untrusted request content before semantic validation. This allowed a remote party with access to the ZMQ endpoint to inject newline and control-character content into daemon logs, enabling log forging. The...

Monero: Authenticated `unsigned_txset` change spoof lets a malicious hot wallet steal cold-signer change

A vulnerability was discovered in Monero's wallet code that allowed a malicious or compromised hot/watch-only wallet with the victim wallet's private view key to forge an authenticated unsigned transaction. The forged transaction included an attacker-controlled output that was presented as change...

Git argument injection in deployment pull steps via unsanitized commit_sha enables RCE on workers

Description Prefect's GitRepository storage class passes user-controlled commitsha values as positional arguments to git commands without a -- separator. The commitsha parameter accepts arbitrary strings with no format validation no hex check, no length check. This allows injection of git flags...