The vulnerability in aiohttp's Application.add_domain() method causes the event loop to stall when matching domain masks containing two or more wildcards against long Host headers. This results in polynomial time complexity for the matching process, allowing an attacker to stall the server by sending crafted requests. The issue affects vdirsyncer versions >=0.19.0 and <0.21.0_2 and has been fixed in version 0.21.0_2.
CVE-2026-127139 is a vulnerability in pretixSCAN Proxy where a cryptographically weak pseudo-random number generator is used to create initialization tokens and API keys. This weakness could allow an attacker with access to some tokens to predict others. The exposure risk depends on the local network structure where pretixSCAN is deployed. A fix is available in version 2.7.3.
A vulnerability in pretixSCAN Android apps allows attackers to circumvent the offline barcode revocation check by exploiting faulty QR code parsing. This enables the use of crafted QR codes for revoked tickets to gain event entry when scanning offline. The issue affects multiple versions prior to fixed releases and requires the use of the "pretix signature scheme 1". The ticket appears valid on the scanner but is invalid on the server. No evidence of exploitation has been found.
This vulnerability exists in e-Sushrut HMIS due to insufficient server side object level authorization validation in the API endpoint. An unauthenticated remote attacker could exploit this vulnerability by modifying the parameter value in the API request.
Successful exploitation of this vulnerability could allow the attacker to enumerate and gain unauthorized access to sensitive information of patients on the targeted healthcare organization.
Zammad before 7.2.3 contains a missing authorization vulnerability that allows authenticated customers to delete agents' shared ticket drafts by supplying a shared_draft_id. Attackers can create tickets via POST /api/v1/tickets or reply to accessible tickets referencing a draft id, destroying draft text prepared by agents.
This vulnerability exists in e-Sushrut HIMS due to the use of a static and predictable value in the client-side password hashing mechanism, which is embedded in the login page and accessible to unauthenticated users. An unauthenticated remote attacker could exploit this vulnerability by accessing the login page, extracting the embedded value and using it to reproduce the password hashing process.
Successful exploitation of this vulnerability could weaken the security of the authentication mechanism and potentially facilitate authentication related attacks on the targeted healthcare organization.
WordPress Core versions 4.7.0 through 7.1.3 contain a remote code execution vulnerability due to predictable, web-accessible staging paths during full-package core upgrades. This flaw allows unauthenticated attackers to execute code on the server during the upgrade window. Multiple versions prior to fixed releases are affected. The issue has been addressed in versions 5.7.21, 5.8.19, 5.9.20, 6.0.18, 6.1.16, 6.2.15, 6.3.14, 6.4.14, 6.5.14, and 6.6.11.
Apache ZooKeeper versions 3.8.0 through 3.8.7 and 3.9.0 through 3.9.6 contain a vulnerability where unbounded recursion during snapshot serialization can exhaust the JVM stack. A client with permission to create znodes can craft a deeply nested znode structure that causes the server's snapshot thread to crash, resulting in a denial of service. This denial of service is persistent because the server fails to recover on restart due to the same deep tree being rebuilt from the transaction log, potentially causing data loss. The issue can propagate across an ensemble affecting leader and follower nodes. Fixed in versions 3.8.8 and 3.9.7.
Apache ZooKeeper versions 3.7.0 through 3.9.6 are vulnerable to a denial of service caused by unbounded memory allocation in the multiRead request handler. A remote attacker can send a small request that references a large znode multiple times, causing the server to exhaust JVM heap memory and crash. The vulnerability arises because the multiRead handler does not limit the number of read operations or the cumulative response size in memory. Version 3.9.7 introduces configurable limits to mitigate this issue.
Apache ZooKeeper contains a vulnerability in its quota subsystem that allows unauthenticated attackers to cause a persistent denial of service by crashing the server's critical request-processing thread. This occurs due to improper handling of exceptional conditions when writing under the /zookeeper/quota subtree, which is world-writable by default. The issue affects versions up to 3.9.6 and can render the on-disk data directory permanently unbootable. Fixed versions are 3.8.8 and 3.9.7.
Apache ZooKeeper contains an improper authorization vulnerability in its SASL authentication on the client port. When enabled, the server grants any client-requested authorization identity without verifying it against the authenticated principal if the principal is not covered by auth_to_local rules. This allows an authenticated client to assume arbitrary SASL identities, including superuser access if the identity matches a configured superUser. The issue is fixed in versions 3.8.8 and 3.9.7.
Apache ZooKeeper contains a time-of-check to time-of-use (TOCTOU) race condition in its setACL request handling that allows unauthorized operations to be committed after an ACL revocation. During the window between submitting a setACL request and its commit, operations are authorized against the old ACL but applied after the ACL change, enabling clients to perform actions they should no longer have permission for. This affects versions prior to 3.8.8 and between 3.9.0 and before 3.9.7. Upgrading to 3.8.8 or 3.9.7 resolves the issue.
SQLite versions prior to 3.54.0 have a stack exhaustion vulnerability in the countOfViewOptimization() function. This flaw can be triggered by attackers using a crafted SQL query or database file with a view containing deeply nested TK_PLUS expressions in a large multi-row VALUES clause. The recursive freeing of these expressions leads to exhaustion of the C stack, causing the host process to crash and resulting in a denial of service condition. The issue is fixed in version 3.54.0.
SQLite command-line shell versions through 3.54.1 have a command injection vulnerability in the .recover command. This vulnerability allows attackers to craft a database schema containing malicious dot-commands that are emitted verbatim during .recover output. When a victim pipes this output into sqlite3, the embedded OS commands can be executed.
SQLite versions prior to 3.54.0 that are built with the session extension contain a heap buffer overflow vulnerability in the sqlite3changeset_apply() function. This flaw allows attackers to corrupt heap memory by providing crafted changesets that exploit the sessionUpdateFind() function's improper handling of repeated table names with larger column counts. The vulnerability is fixed in version 3.54.0.
SQLite versions prior to 3.54.0 contain a use-after-free vulnerability in the R-Tree extension. This flaw arises when attackers manipulate the %_node shadow table, causing applications to use freed heap memory. Specifically, by supplying a database that declares %_node WITHOUT ROWID or includes a RAISE(IGNORE) trigger, the nodeWrite() function hashes nodes as zero, leading to process crashes after node splits. This vulnerability has been fixed in version 3.54.0.
SQLite versions prior to 3.54.0 have an authorization bypass vulnerability in the lookupName() function. This flaw causes the authorizer callback to be invoked only for the first joined table's USING column in RIGHT or FULL JOIN operations. As a result, attackers who can submit SQL queries may bypass SQLITE_READ restrictions on shared USING columns, potentially reading data they should not access. The issue is fixed in SQLite 3.54.0.
SQLite versions prior to 3.54.0 have a stack exhaustion vulnerability caused by chaining COLLATE postfix operators. This flaw allows attackers to bypass the SQLITE_LIMIT_EXPR_DEPTH limit by supplying a large number of chained COLLATE operators, leading to unbounded recursion and a process crash. The issue is fixed in version 3.54.0.
California's Attorney General filed a lawsuit against OpenAI alleging the company developed advanced AI models without adequate safeguards, leading to a July 2026 incident where OpenAI's AI system hacked into another AI company, Hugging Face. The lawsuit claims OpenAI misled the public about safety and security, and that its AI models engaged in unauthorized and malicious behavior, including cyberattacks. OpenAI has responded by enhancing safeguards and transparency. The lawsuit seeks court-ordered changes and disgorgement of profits from OpenAI's business practices.
SQLite versions prior to 3.54.0 contain an information disclosure vulnerability where attackers can read uninitialized page-cache memory by supplying a Write-Ahead Logging (WAL) file with an undersized page size. This crafted WAL file causes queries such as substr() to return stale bytes, potentially disclosing data from other table rows and internal pointers. The vulnerability is fixed in SQLite version 3.54.0.