CVE-2026-56446: Misp-Project Misp Code Injection (CVSS 8.7)

Affected versions of MISP permit unauthenticated or weakly constrained request paths to perform persistent work without adequate input bounds or rate limiting. The users/forgot password-reset endpoint accepted an attacker-controlled email value without first imposing a reasonable length bound or validating its format. That value was then used to create an audit log entry and queue a password-reset job, causing the supplied value to be persisted more than once per request. The commit explicitly states that an unbounded unauthenticated request field was stored twice per call with no throttle. The fix adds:  - a maximum email input length of 1024 bytes;  - email-format validation before persistent work;  - a per-source pre-authentication request budget;  - HTTP 429 responses when that budget is exceeded;  - a 15-minute cooldown for API-access request emails;  - POST-only handling and CSRF protection for the API-access request endpoint. The new flood filter is specifically intended to limit persistent storage costs from anonymous requests such as password resets, registrations, and failed REST authentication attempts. Version affected: ≤2.5.45

Affected versions of MISP contain insufficient validation of server-side outbound HTTP destinations in feed retrieval and TAXII discovery functionality. In feed processing, redirects were followed without validating the redirect scheme or destination. The original request headers were reused across redirect hops, meaning authentication headers or API credentials configured for a feed could be forwarded to a different host. Redirects could also target internal network resources, resulting in SSRF. The fix adds redirect validation, blocks internal destinations for cross-host redirects, strips configured feed credentials before following redirects to another host, and pins validated DNS results to prevent re-resolution after validation. The TAXII discovery endpoint had a related incomplete SSRF defense. It used gethostbyname() and compared the result against only a few literal addresses. This missed cases including IPv6 loopback (::1), numeric host encodings such as 0x7f000001, and potentially multiple DNS records. The fix moves TAXII discovery to the shared URL egress validator. Together, these commits harden MISP's outbound URL handling against alternate-address representations, DNS-related bypasses, unsafe redirects, internal-host access, and cross-host credential forwarding. Version affected: ≤2.5.45

A vulnerability in MISP's email-based one-time password (OTP) authentication flow allowed an attacker to perform an unrestricted number of OTP verification attempts. The email_otp() endpoint did not apply brute-force protection when validating submitted OTP values. An attacker who had reached the OTP verification stage, for example after successfully providing a user's primary authentication credentials, could repeatedly submit candidate OTP values while the same OTP remained valid. This significantly increased the feasibility of guessing the OTP and bypassing the additional authentication factor, potentially resulting in unauthorized access to the affected user's account. The issue was exacerbated by the fact that the OTP is associated with the user rather than with an individual pending login session, allowing multiple concurrent sessions to attempt guesses against the same valid OTP. The patch integrates the existing MISP brute-force protection mechanism into the email OTP flow. Failed OTP attempts are now counted against the user, further attempts are rejected once the configured threshold is reached, and the active OTP is invalidated when the attempt budget is exhausted. Blocklisted users are also prevented from requesting the generation of a fresh OTP. In addition, OTP comparison now uses hash_equals() and validates that the submitted value is a string.

A cross-site request forgery (CSRF) vulnerability existed in the cullEmptyEvents action of MISP. The endpoint performed a state-changing and irreversible operation while accepting HTTP GET requests. Because bodyless GET requests are not subject to CakePHP's CSRF validation, an attacker could cause an authenticated MISP user with sufficient privileges to invoke the endpoint simply by causing their browser to load a crafted URL, for example through an embedded image or other automatically requested resource. Successful exploitation triggers the deletion of published empty events. The deletion is particularly significant because the operation uses skipBlocklist, meaning the removed events do not leave blocklist entries that could prevent or track their subsequent synchronization. This can result in unintended and potentially irreversible deletion of MISP event records without explicit user interaction. The vulnerability was addressed by restricting cullEmptyEvents to HTTP POST requests, ensuring that CakePHP's normal CSRF protections are applied to the operation.

MISP contains an improper TLS certificate validation vulnerability in CurlClient. The CurlClient::$verifyPeer property was not explicitly initialized and therefore defaulted to null. When passed to cURL, this value effectively disabled TLS peer verification unless the calling code explicitly enabled it. As a result, HTTPS connections made through affected CurlClient instances could accept certificates that were not issued by a trusted certificate authority. An attacker capable of intercepting or manipulating network traffic between a MISP instance and a remote HTTPS service could impersonate the remote endpoint and perform a man-in-the-middle attack. Successful exploitation could allow an attacker to observe sensitive information transmitted by MISP, including authentication material or exchanged threat intelligence, and to modify responses returned to the MISP instance. The impact depends on the functionality using CurlClient and the data exchanged with the remote service. The patch enables TLS peer verification by default while preserving explicit support for configured self-signed certificates. It also corrects the self-signed certificate handling in SyncTool so that peer verification is disabled only when no pinned CA certificate is configured.

MISP contains an authentication bypass vulnerability in its LDAP and LinOTP authentication components due to insufficient validation of user-supplied credentials. The custom LdapAuthenticate and LinOTPAuthenticate components replace CakePHP's FormAuthenticate implementation but did not replicate its credential validation checks. As a result, empty or non-string values could reach the underlying authentication mechanisms. In the LDAP authentication path, an attacker able to identify a valid directory user's email address could submit an empty password. The empty credential could be passed to ldap_bind(), where an LDAP server accepting unauthenticated binds may return a successful result for a valid distinguished name combined with an empty password. MISP could consequently treat the attacker as the corresponding authenticated directory user without verification of the user's password. The issue also affected the LinOTP authentication component. Invalid credential types were not rejected before being processed, and when mixed authentication was enabled, an empty password could be checked against a locally stored MISP password hash. LDAP-provisioned MISP accounts could additionally be created with an empty local password because account creation skipped normal validation, resulting in a hash corresponding to an empty password. This could permit authentication through the local fallback mechanism when such an account was no longer resolved through LDAP. Successful exploitation could allow a remote unauthenticated attacker to impersonate an existing MISP user. If the targeted account has administrative or other privileged permissions, the attacker could gain corresponding access to sensitive threat-intelligence data, modify or delete information, alter configuration, or perform other privileged operations. The patch resolves the vulnerability by requiring authentication identifiers and passwords to be valid strings, rejecting empty passwords where they are not explicitly permitted, and assigning a randomly generated local password to LDAP-provisioned accounts instead of storing a hash derived from an empty password.