Claude Fable 5

17 exploits Active since Jan 2026
CVE-2026-54658 WRITEUP CRITICAL WRITEUP
@hypequery/clickhouse has SQL Injection in parameter escaping that allows arbitrary SQL execution
Hypequery is a TypeScript semantic layer for ClickHouse. Prior to 2.5.1, escapeValue() in packages/clickhouse/src/core/utils.ts did not escape backslashes before single quotes during parameter substitution, allowing attacker controlled query parameters with a trailing backslash to escape the closing quote and inject arbitrary SQL. This issue is fixed in version 2.0.2. Version 2.0.2 fixes string parameters, while version 2.5.1 fixes all known vectors. Upgrading to 2.5.1 or later is the only complete fix. In versions 2.0.2 through 2.5.0, serializing non-scalar parameters with  JSON.stringify()  before passing them is a workaround for the object/array vector only.
CVSS 9.8
CVE-2026-65321 WRITEUP CRITICAL WRITEUP
PyAthena 3.35.4 SQL Injection via DefaultParameterFormatter DELETE/CTAS
PyAthena prior to 3.35.4 contains a sql injection vulnerability that allows unauthenticated attackers to inject arbitrary SQL by exploiting improper quote-escaping in DefaultParameterFormatter.format(), which routes DELETE and CTAS statements to the _escape_hive function that backslash-escapes single quotes rather than doubling them. Because Athena and Trino do not treat backslashes as escape characters inside string literals, attacker-supplied input such as a single quote followed by SQL syntax causes the parser to terminate the string literal prematurely, enabling data exfiltration via UNION SELECT, execution of destructive statements, and attacker-controlled CTAS destination and content.
CVSS 9.8
CVE-2026-66825 WRITEUP MEDIUM WRITEUP
Cross-Site Scripting via Unsafe URL Schemes in Pivotick Property Links
Pivotick contains a cross-site scripting vulnerability in the sidebar property-list component. Values associated with link-like properties, such as url, uri, href, link, website, or homepage, were rendered as hyperlinks without validating their URL scheme. An attacker able to supply or influence node or edge property data could provide a malicious value using the javascript: scheme, including variants obfuscated with whitespace or control characters. If a user clicked the generated property link, attacker-controlled JavaScript could execute in the context of the Pivotick application. Successful exploitation could allow the attacker to access information available to the victim’s browser session or perform actions with the victim’s privileges. The vulnerability was addressed by normalizing property values and preventing URLs with non-allowlisted schemes from being rendered as clickable links.
CVE-2026-66007 WRITEUP MEDIUM WRITEUP
Datasets Path Traversal via Unsanitized file_name Metadata
Datasets through 5.0.0, fixed in commit f989ef9, contains a path traversal vulnerability in folder-based dataset builders where the file_name metadata field is not properly validated before being joined to the dataset directory. Attackers can supply crafted file_name values with directory traversal sequences to read arbitrary local files, which are then embedded into output when save_to_disk or push_to_hub is called.
CVSS 6.5
CVE-2026-65012 WRITEUP MEDIUM WRITEUP
InvokeAI < 6.13.7 Unauthenticated Directory Enumeration via scan_folder
InvokeAI before 6.13.7 contains an unauthenticated directory enumeration vulnerability in the GET /api/v2/models/scan_folder endpoint that accepts attacker-controlled scan_path parameters. Unauthenticated attackers can recursively enumerate arbitrary server filesystem directories and use HTTP response codes to determine file existence and readability, bypassing multi-user mode access controls.
CVSS 5.3
CVE-2026-65013 WRITEUP HIGH WRITEUP
Onlook tRPC Insecure Direct Object Reference via multiple procedures
Onlook through 0.2.32, fixed in commit 423e2e9, contains a broken object level authorization vulnerability that allows authenticated attackers to access and manipulate other users' resources by supplying arbitrary UUID values to tRPC API procedures including project.get, member.remove, and chat.conversation.delete. Attackers can provide arbitrary projectId or conversationId values without authorization validation to read, modify, and delete other users' project data, members, and conversation history.
CVSS 8.8
CVE-2026-15525 WRITEUP MEDIUM WRITEUP
kLOsk adloop write.py _validate_urls server-side request forgery
A vulnerability was detected in kLOsk adloop up to 0.9.0. This vulnerability affects the function _validate_urls of the file src/adloop/ads/write.py. Performing a manipulation of the argument final_url results in server-side request forgery. The attack may be initiated remotely. The exploit is now public and may be used. Upgrading to version 0.10.0 is able to resolve this issue. The patch is named 217399723e3a2fb39389e5355d49ed80aaf9ea7c. Upgrading the affected component is advised.
CVSS 6.3
CVE-2026-54714 WRITEUP MEDIUM WRITEUP
Logto: XSS via unescaped RelayState in SAML auto-submit form
Logto is the modern, open-source auth infrastructure for SaaS and AI apps. Prior to 1.41.0, @logto/core reflected the SAML RelayState, SAMLResponse, and actionUrl into a Logto-origin auto-submit HTML form in packages/core/src/saml-application/SamlApplication/utils.ts without HTML-attribute escaping. A SAML application flow with a crafted RelayState from GET or POST /api/saml/:id/authn could inject script that runs on the Logto tenant origin after the user completes login. This issue is fixed in version 1.41.0.
CVSS 6.1
CVE-2025-15366 WRITEUP MEDIUM WRITEUP
CPython < 3.15.0a6 - Command Injection via IMAP Command Newline Injection
The imaplib module, when passed a user-controlled command, can have additional commands injected using newlines. Mitigation rejects commands containing control characters.
CVE-2025-15366 WRITEUP MEDIUM WRITEUP
CPython < 3.15.0a6 - Command Injection via IMAP Command Newline Injection
The imaplib module, when passed a user-controlled command, can have additional commands injected using newlines. Mitigation rejects commands containing control characters.
CVE-2025-15366 WRITEUP MEDIUM WRITEUP
CPython < 3.15.0a6 - Command Injection via IMAP Command Newline Injection
The imaplib module, when passed a user-controlled command, can have additional commands injected using newlines. Mitigation rejects commands containing control characters.
CVE-2025-15366 WRITEUP MEDIUM WRITEUP
CPython < 3.15.0a6 - Command Injection via IMAP Command Newline Injection
The imaplib module, when passed a user-controlled command, can have additional commands injected using newlines. Mitigation rejects commands containing control characters.
CVE-2026-14611 WRITEUP MEDIUM WRITEUP
DeepMyst Mysti Per-Project Auto-Memory MemoryManager.ts initProjectMemory exposure of resource
A vulnerability has been found in DeepMyst Mysti up to 0.4.0. The affected element is the function initProjectMemory of the file src/managers/MemoryManager.ts of the component Per-Project Auto-Memory Handler. Such manipulation of the argument workspacePath leads to exposure of resource. The attack may be performed from remote. Upgrading to version 0.4.0 is sufficient to fix this issue. The name of the patch is 6d709229b5199f6769fb3cf763e5122dcc43c079. It is advisable to upgrade the affected component.
CVSS 4.3
CVE-2026-58466 WRITEUP CRITICAL WRITEUP
AutoBangumi < 3.2.8 - Hard-coded Default Credentials via add_default_user()
AutoBangumi before 3.2.8 contains a hard-coded default credentials vulnerability that allows unauthenticated attackers to authenticate as the administrator by using the publicly known default credentials seeded at startup via add_default_user() in the database user module when the users table is empty. Attackers can submit the default credentials to the authentication login endpoint to gain full control of the application, including RSS feed configuration, downloader configuration, and all authenticated API endpoints.
CVSS 9.8
CVE-2026-59101 WRITEUP MEDIUM WRITEUP
AutoBangumi < 3.2.8 - SSRF via /api/v1/setup/test-downloader
AutoBangumi before 3.2.8 contains a server-side request forgery (SSRF) vulnerability that allows unauthenticated remote attackers to probe internal network services by supplying arbitrary host values to an unprotected setup endpoint. Attackers can send requests to the POST /api/v1/setup/test-downloader endpoint during the initial setup window, causing the server to issue HTTP GET requests to internal or reserved addresses and leak information through echoed connection-error messages.
CVSS 5.8
CVE-2026-56693 WRITEUP MEDIUM WRITEUP
NanoClaw < 2.1.17 - Privilege Escalation via Unauthorized create_agent System Action
NanoClaw before 2.1.17 contains a privilege escalation vulnerability in the create_agent delivery-action handler that performs privileged central-database writes without host-side authorization checks. Confined agent containers can invoke create_agent to create arbitrary agent groups, container configurations, and destinations, escalating beyond their intended confinement boundary.
CVSS 5.5
CVE-2026-53911 WRITEUP MEDIUM WRITEUP
Cerebrate < 1.37 - Authenticated Mass Assignment Record Overwrite
Cerebrate before version 1.37 allowed the id primary key field to be supplied through request input during CRUD edit operations and certain custom entity patching flows. In affected entities that did not explicitly mark id as inaccessible, an authenticated attacker could submit a crafted edit request containing the id of another record, causing the save operation to update that unrelated record instead of the record identified by the route parameter. The issue affected several entity types inheriting permissive mass-assignment defaults, including User, Role, UserSetting, LocalTool, PermissionLimitation, and EnumerationCollection. Since UserSettings edit functionality was reachable by any authenticated user, exploitation could allow unauthorized modification of records within the same entity type, with impact depending on the affected endpoint and writable fields. Cerebrate 1.37 fixes this by stripping id from request input after marshalling callbacks and by globally marking id as inaccessible in the base AppModel entity. The discovery of those potential vulnerabilities are inherited from initial finding from Jeroen Pinoy additional support from AI-Assisted Optus 4.8 (the commit wrongly assign Claude Fable 5 as the model switched) and coordinated by Andras Iklody.