The vulnerability arises from misp-stix's STIX import logic automatically choosing between internal and external parsers based on metadata in the STIX document, which is fully controlled by the document producer and thus untrusted. For STIX2, MISP-specific tool labels could cause misclassification as MISP-origin content; for STIX1, the document title was used similarly. When STIX2 content was processed as an internal MISP export, the entire x_misp_attributes dictionary was copied without restricting fields, allowing crafted STIX bundles to inject security-sensitive attribute properties like distribution, sharing_group_id, and tags. This could lead to unauthorized sharing or influence on automated processing. The root cause is dynamically assigning externally supplied object properties without an allow-list (CWE-915) and using untrusted values for security classification decisions (CWE-807). The attack aligns with CAPEC-153, involving input data manipulation to alter processing paths.

An attacker able to supply a crafted STIX document for import can manipulate MISP attribute metadata, including distribution and sharing restrictions, potentially causing information to be shared against organizational policies or influencing downstream automation based on attacker-controlled metadata. This could undermine data confidentiality and integrity within the MISP environment.

Patch status is not yet confirmed — check the vendor advisory for current remediation guidance. The disclosed fix introduces an explicit classification parameter to override parser selection based on untrusted metadata and implements an allow-list restricting which attribute properties can be set during import, addressing the root causes. Until a patch is applied, organizations should exercise caution when importing STIX documents from untrusted sources.