Total
395716 CVE
| CVE | Vendors | Products | Updated | CVSS v2 | CVSS v3 |
|---|---|---|---|---|---|
| CVE-2026-57131 | 2026-09-16 | N/A | 9.8 CRITICAL | ||
| PraisonAI is a multi-agent teams system. Prior to 4.6.58, praisonai.jobs.server.create_app mounts praisonai.jobs.router.create_router under /api/v1/runs without authentication or per-job authorization. Network clients can submit attacker-controlled prompts and agent configuration, list and read jobs, stream results, and cancel or delete other jobs, exposing service credentials and connected tool capabilities to unauthorized agent execution. This vulnerability is fixed in 4.6.58. | |||||
| CVE-2026-57127 | 2026-09-16 | N/A | 9.8 CRITICAL | ||
| PraisonAI is a multi-agent teams system. Prior to 4.6.58, recipe serve installs APIKeyAuthMiddleware or JWTAuthMiddleware when an operator selects api-key or JWT authentication, but each middleware forwards requests when PRAISONAI_API_KEY or PRAISONAI_JWT_SECRET and the corresponding recipe value are absent. Unauthenticated clients can then reach recipe execution, input, and output surfaces and may trigger connected tools despite the operator explicitly enabling authentication. This issue is fixed in 4.6.58. | |||||
| CVE-2026-56839 | 2026-09-16 | N/A | 7.3 HIGH | ||
| PraisonAI is a multi-agent teams system. Prior to 4.6.59, the CODE_TOOLS wrappers keep _workspace_root as None and pass workspace=None to read_file, search_replace, and apply_diff helpers that enforce path containment only for a truthy workspace. An application that exposes code_read_file, code_search_replace, or code_apply_diff before set_workspace can therefore let prompt-influenced calls read and modify files outside the intended project directory, while explicitly configured workspaces remain effective. This vulnerability is fixed in 4.6.59. | |||||
| CVE-2026-90961 | 2026-09-16 | N/A | N/A | ||
| The LdapAuth and LinOTPAuth authentication plugins in MISP contain an authentication bypass vulnerability. Both LdapAuthenticate and LinOTPAuthenticate replace CakePHP's FormAuthenticate class but fail to replicate its _checkFields() input validation guard. As a result, the email and password fields extracted from the login request are passed to downstream authentication logic without verifying that they are non-empty strings. In the LDAP authenticator, an empty or null password is forwarded to ldap_bind(). Per RFC 4513 section 5.1.2, a bind request with a valid DN and an empty password constitutes an unauthenticated bind, which many LDAP directory servers accept as successful. An attacker who knows any valid user email address in the directory can therefore authenticate as that user without possessing a password. Additionally, non-string values (null, false, arrays) are either coerced to empty strings by ldap_bind(), raise TypeErrors, or are misinterpreted as find conditions in _findUser(), all of which can lead to unintended authentication outcomes. In the LinOTP authenticator, the same missing guard allows non-string credentials to be concatenated into the LinOTP verification request, and in the mixed-authentication branch an empty password is accepted against a stored hash of the empty string. A secondary issue in the LDAP authenticator is that newly created user accounts (auto-provisioned on first LDAP login) were assigned an empty password. Because the save path skips validation, the empty string is hashed and stored. If the user later ceases to be found in LDAP and the mixed-authentication fallback is used, the stored hash of the empty string verifies against an empty password, again permitting unauthenticated access. The vulnerability requires that the affected plugin (LdapAuth or LinOTPAuth) is enabled on the MISP instance and that the attacker knows at least one valid email address registered in the directory or MISP user store. No prior authentication is required. Successful exploitation grants the attacker the full privileges of the impersonated user, which may include administrative access to threat intelligence data. Version affected: ≤2.5.45 | |||||
| CVE-2026-90957 | 2026-09-16 | N/A | N/A | ||
| Affected versions of MISP serve uploaded SVG images inline without a restrictive browser sandbox. The commit explains that SVG files are XML documents rather than passive bitmap images. While scripts inside SVG do not execute when the SVG is rendered through a normal <img>, they can execute when the SVG is navigated to directly or embedded as a document. In that case, malicious <script> elements, event handlers, or javascript: URLs execute on the MISP origin with the viewer’s session. The affected use cases include: - organisation SVG logos; - event-report SVG pictures. Importantly, the vulnerable behavior is on the serve path, not merely the upload path: the patch notes that a malicious SVG uploaded while SVG support was enabled could remain dangerous even after uploads were later disabled. Version affected: ≤2.5.45 | |||||
| CVE-2026-90895 | 2026-09-16 | N/A | N/A | ||
| Affected versions of MISP’s interactive CLI shell implement access control independently from the normal web application, causing several authorization inconsistencies. The patch shows that CLI access could differ from the web application in multiple security-sensitive areas: - feed listings did not enforce the same lookup_visible restrictions for non-host-organisation users; - feed detail access did not enforce the same host-organisation/site-admin authorization as FeedsController::view(); - Feed.headers, which can contain HTTP authorization credentials, could be exposed instead of being hidden or masked; - server synchronization authkey values were not explicitly hidden from CLI detail output; - sharing-group detail access did not consistently use SharingGroup::checkIfAuthorised(); - the use command could establish context for a record without first proving that the user was authorized to view that record The commit additionally hardens pagination and terminal rendering, including neutralization of terminal control sequences found in database-backed values. Those are important hardening changes, but the main vulnerability is the CLI authorization/data-disclosure mismatch. Version affected: ≤2.5.45 | |||||
| CVE-2026-90893 | 2026-09-16 | N/A | N/A | ||
| MISP contains a Cross-Site Request Forgery (CSRF) vulnerability in the UserSettingsController. The actions setTheme, setHomePage, and eventIndexColumnToggle were explicitly added to the Security component's unlockedActions list, which disabled all CSRF validation (both token and field-hash checks) for those endpoints. Because these endpoints accept POST requests and modify per-user application state (theme selection, default homepage URL, and event index column visibility), an attacker who can induce a logged-in MISP user to load a malicious page (e.g., via a crafted link, embedded image, or auto-submitting form) can forge requests that alter the victim's settings without their knowledge or consent. The most impactful action is setHomePage, which allows an attacker to redirect the victim's default landing page to an arbitrary attacker-controlled URL, potentially facilitating phishing or further social engineering. The setTheme action can alter the user's visual theme, and eventIndexColumnToggle can change which columns are displayed in the event index view. No authentication bypass is involved; the victim must already be authenticated to MISP. The vulnerability was reported by the Scottish Government National Cyber Team. Version affected: ≤2.5.45 | |||||
| CVE-2026-29812 | 2026-09-16 | N/A | 4.3 MEDIUM | ||
| CyberPanel before 2.4.4 has no logging for actions that could potentially manipulate the child domains list. | |||||
| CVE-2026-29810 | 2026-09-16 | N/A | 4.3 MEDIUM | ||
| CyberPanel before 2.4.4 omits a "return 0" that is required by the business logic. | |||||
| CVE-2026-90955 | 2026-09-16 | N/A | N/A | ||
| Affected versions of MISP’s interactive CLI shell do not reliably preserve the identity of the impersonated MISP user across audit logging. The shell is designed to run actions as a supplied MISP user ID. However, the legacy SysLogLogable behavior stored that identity in behavior-instance state that could be overwritten when another model lazily attached the shared behavior. Consequently, subsequent CLI writes could lose the intended user attribution and be logged incorrectly. The commit also notes that CLI-originated records lacked a CLI marker, making them appear similar to ordinary web actions by that user. Version affected: ≤2.5.45 | |||||
| CVE-2026-86118 | 2026-09-16 | N/A | 4.3 MEDIUM | ||
| gonic versions before 0.22.0 fail to validate administrator privileges in the startScan endpoint, allowing any authenticated user to trigger media library rescans. Attackers can repeatedly call the startScan endpoint to force CPU and I/O-intensive filesystem operations, causing denial of service on multi-user instances. | |||||
| CVE-2026-49838 | 2026-09-16 | N/A | 5.9 MEDIUM | ||
| GoBGP is an open source Border Gateway Protocol (BGP) implementation in the Go Programming Language. Prior to version 4.7.0, GoBGP accepts a zero-length AS_PATH during UPDATE decoding and later panics while validating that attribute for a confederation eBGP peer. The vulnerable path is in the BGP UPDATE validator: a malformed UPDATE that should be rejected as a malformed AS_PATH instead reaches an unchecked `p.Value[0]` access, allowing a configured confederation eBGP peer to trigger a denial of service. Version 4.7.0 patches the issue. | |||||
| CVE-2026-49837 | 2026-09-16 | N/A | 5.9 MEDIUM | ||
| GoBGP is an open source Border Gateway Protocol (BGP) implementation in the Go Programming Language. Versions prior to 4.6.0 contain a BGP OPEN capability parsing issue where several concrete capability decoders may parse data from the full remaining capability buffer instead of the slice bounded by the declared capability length, `CapLen`. A malformed BGP OPEN message can cause bytes from a following capability to be interpreted as part of the current capability. The most security-relevant case is the 4-octet AS capability, where a capability with `CapLen == 0` may cause the parser to read bytes from the following capability as the 4-octet AS value. This parsed value may later affect peer AS validation during BGP session establishment. Version 4.6.0 patches the issue. | |||||
| CVE-2026-81029 | 2026-09-16 | N/A | 8.1 HIGH | ||
| OpenMetadata accepts a caller-supplied post-authentication redirect target and appends the issued token to it. SamlLoginServlet reads the callback request parameter and stores it in the HTTP session without comparing it against any configured or registered destination, and the assertion consumer servlet later formats that stored value into a URL carrying the freshly issued JWT together with the account's email and name before sending the redirect. The OIDC and OAuth2 handler follows the same pattern with its own redirect parameter and the issued identity token. A request naming a destination the attacker controls therefore causes the server to deliver a valid token for whoever completes the login to that destination. Because the token authenticates API calls as that account, a user who follows such a link and authenticates hands over control of their account. Version 2.0.0 removes the caller-supplied callback parameter; no 1.x release validates it. | |||||
| CVE-2026-29811 | 2026-09-16 | N/A | 7.7 HIGH | ||
| CyberPanel before 2.4.4 attempts to detect an "alais" domain (i.e., a second domain that serves the same content as a primary domain; normally spelled "alias") via an ORM query filter rather than a Python "if" statement. | |||||
| CVE-2025-64059 | 2026-09-16 | N/A | 1.8 LOW | ||
| Grav 1.7.50.2 allows admins to enter JavaScript via the Home Page editor. NOTE: the relevance of this for stored XSS is disputed because admins are allowed to modify templates, install plugins, and upload other executable content. | |||||
| CVE-2026-86174 | 2026-09-16 | N/A | 4.3 MEDIUM | ||
| Plane through 1.4.2 fails to validate that issues belong to the deploy board's project in the public comment endpoint. Authenticated attackers can post comments to arbitrary issues across workspaces by supplying an issue_id parameter to the public deploy-board comment endpoint. | |||||
| CVE-2026-56100 | 2026-09-16 | N/A | 8.1 HIGH | ||
| SpringBlade versions from 2.7.3 up to but not including 5.0.0 contain a privilege escalation vulnerability that allows authenticated attackers to create system administrator accounts by sending crafted POST requests to an unprotected internal Feign user-creation endpoint exposed via @RestController without authorization checks. Attackers can exploit the gateway's authentication filter, which only validates JWT parsing without verifying user roles or caller identity, and leverage a hardcoded JWT signing key embedded in publicly available JARs to forge tokens and escalate privileges from a low-privilege user to administrator, enabling cross-tenant data pollution and persistent backdoor access. | |||||
| CVE-2026-80350 | 2026-09-16 | N/A | 7.1 HIGH | ||
| OneUptime's webhook target check rejects private and loopback addresses given in IPv4 form and a small set of IPv6 forms, but has no case for the IPv4-mapped IPv6 range. The webhook delivery path calls SSRFProtection.validateWebhookTargetIsSafe, and the host-literal screening inside Common/Server/Utils/SSRFProtection.ts, performed by isBlockedHostnameLiteral, rejects private and loopback IPv4 ranges and tests an IPv6 value against the unspecified address, the loopback, the link-local prefix and the unique-local prefixes. A value such as [::ffff:127.0.0.1] matches none of them. The value is also recognised as an address literal rather than a name, so the path that re-checks addresses obtained from resolution is not taken. The HTTP client treats the mapped form as the embedded IPv4 address and connects to it, so an authenticated project member who can configure a webhook can direct the server at loopback services, private network ranges and link-local metadata endpoints, and the response is recorded where the webhook result can be read. Version 12.0.7 adds handling for the mapped range. | |||||
| CVE-2026-73529 | 2026-09-16 | N/A | 5.3 MEDIUM | ||
| Plainpad through 1.1.1, fixed in commit d3823fc, contains a missing rate limiting vulnerability that allows unauthenticated attackers to send unbounded login requests to the POST /v1/sessions endpoint due to dead code in App\Http\Kernel.php that is never instantiated under the Laravel 11+ skeleton, leaving the API throttle configuration unattached to any route. Attackers can exploit this by sending unlimited credential attempts with no lockout or CAPTCHA enforcement, and additionally trigger CPU exhaustion by forcing repeated bcrypt comparisons on each request. | |||||