CAI Content Credentials CVE-2026-100120

Exploitation requires that the target application embed and expose the CAI Content Credentials SDK's input-processing/validation path to attacker-supplied content or requests over the network; the flaw is an improper-input-validation bypass of the SDK's security control that grants unauthorized write access. Per CVSS AV:N/AC:L/PR:N/UI:N, no authentication and no user interaction are needed against a reachable deployment, so there are few limiting preconditions beyond network reachability of an application using a vulnerable SDK version. The practical limiting factor is footprint: only applications that integrate this specific SDK (and have not updated to the fixed version in APSB26-163) are affected, and impact is confined to integrity (C:N/A:N), i.e., tampering with content-credential/provenance data rather than full compromise.

The CVSS 3.1 vector (AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:H/A:N, base 7.5) describes a network-reachable, low-complexity, unauthenticated flaw with high integrity impact and no confidentiality or availability impact - consistent with the description of a security-feature bypass yielding unauthorized write access without user interaction. No EPSS score, KEV listing, or POC indicator was supplied in the input, so there is no public exploit identified at time of analysis and no evidence of active exploitation; the risk signal rests entirely on the CVSS vector and the Adobe PSIRT advisory. The integrity-only profile means this is a data-integrity/provenance-tampering issue (consistent with a Content Credentials SDK whose purpose is content authenticity) rather than a full system compromise, so it is a moderate real priority: worth patching promptly where the SDK is deployed, but not an emergency absent exploitation data. Note a potential conflict: the input Tag labels this 'Authentication Bypass,' while the description and CWE-20 characterize it as an improper-input-validation security-feature bypass granting write access, not necessarily an authentication bypass - verify the precise bypassed control with Adobe advisory APSB26-163. EPSS is absent (not an error), so widespread-versus-targeted exploitation likelihood cannot be estimated from the data provided.

A remote, unauthenticated attacker sends crafted input to an application using the CAI Content Credentials SDK that is not properly validated, bypassing the SDK's integrity/security checks to gain unauthorized write access and tamper with or forge content-credential (provenance) data. Because exploitation requires no user interaction and no privileges (AV:N/AC:L/PR:N/UI:N), it can be automated against reachable endpoints. No public exploit code was identified in the provided data.