Content Credentials CVE-2026-48290

Exploitation requires the victim to use an application that embeds an affected Adobe C2PA component (c2pa Rust SDK ≤0.84.0, @contentauth/c2pa-web ≤0.7.0, or c2patool ≤0.16.5) AND to actively interact - per CVSS UI:R the victim must visit a maliciously crafted URL or interact with a compromised web page that supplies attacker-controlled Content Credentials content for validation. The specific prerequisite is that the affected SDK validates an untrusted manifest whose reference the attacker controls, triggering the SSRF fetch. Limiting factors: the vector is AV:L (no directly network-reachable server endpoint) and mandatory user interaction, so exploitation is not automatable (SSVC Automatable: no) and cannot occur against a target that never opens attacker-supplied Content Credentials data.

The signals point to high technical severity but low current exploitation pressure. The CVSS 3.1 base of 8.2 (High) is driven by S:C, C:H, and I:H, and CISA SSVC rates Technical Impact as 'total' - a successful exploit fully compromises the affected user context. However, the vector is AV:L (local attack vector) requiring UI:R (the victim must open a crafted URL or interact with a compromised page), which raises the practical bar. Corroborating low immediacy: EPSS is only 0.18% (7th percentile), SSVC Exploitation is 'none', SSVC Automatable is 'no', and the flaw is not in CISA KEV - no evidence of active or scalable exploitation. One tension worth flagging: SSRF (CWE-918) is conventionally a server-side issue, yet the assigned vector is AV:L with UI:R, reflecting that exploitation runs through a user-side verification/rendering flow rather than a remotely reachable server endpoint - verify the exact trigger surface against Adobe advisory APSB26-80 before assuming network reachability. Overall this is a genuinely high-impact but low-immediacy issue: patch on a normal cycle and prioritize where these SDKs process untrusted user-supplied media.

An attacker publishes a media file or web page carrying a crafted Content Credentials manifest and lures a victim into opening it in an application built on an affected C2PA SDK. When the SDK validates the manifest it follows the attacker-controlled reference (SSRF), which is chained into script injection and code execution in the victim's session, allowing account or session takeover. No public exploit code has been identified, and the AV:L/UI:R vector means the attack succeeds only after the victim's interaction.