CVE-2026-47904: CAI Content Credentials Resource Exhaustion Denial-of-Service

Technical summary

CVE-2026-47904 is an Uncontrolled Resource Consumption vulnerability (CWE-400) affecting Adobe's c2pa-web and c2pa libraries at version 0.7.1 and earlier for c2pa-web, and v0.80.1 and earlier for c2pa respectively. The vulnerability allows an unauthenticated, local attacker to exhaust CPU, memory, or other critical system resources, leading to a denial-of-service condition. No user interaction is required; the vulnerability can be triggered through crafted input or direct API calls. The CVSS v3.1 score of 6.2 (Medium) reflects that while availability impact is high, the attack vector is localized and requires local system access.

Business impact

Organizations embedding CAI Content Credentials in applications face operational disruption risk. A successful exploitation could render authentication or content verification services unavailable, disrupting workflows that depend on digital content provenance tracking. While the attack surface is limited to local attackers, in multi-tenant or cloud-deployed scenarios, a compromised or malicious user account could impact service availability for all tenants. For media, publishing, and enterprise organizations relying on CAI for authenticity verification, downtime directly affects content trust operations.

Affected systems

The vulnerability affects Adobe's Content Credentials libraries: c2pa-web versions 0.7.1 and earlier, and c2pa versions 0.80.1 and earlier. Both are actively used in web applications and services for embedding content authenticity claims. Any application or service bundling these versions is in scope. Organizations should inventory deployments of these libraries, particularly in production environments, and note their current pinned versions.

Exploitability

Exploitation requires local system access; remote attackers cannot directly trigger this vulnerability over a network. The attack has low complexity and requires no elevated privileges or user interaction, making it straightforward for any local process or authenticated user to execute. The barrier to exploitation is modest once system access is established, but the attack surface is confined to local threat actors.

Remediation

Upgrade c2pa-web to a version released after 0.7.1 and c2pa to a version released after 0.80.1. Verify the specific patched versions in Adobe's security advisory before deployment. Additionally, implement resource quotas and monitoring on systems running these libraries to detect and mitigate resource exhaustion patterns in real time.

Patch guidance

Check Adobe's security advisory for the exact patched versions of both c2pa and c2pa-web released after the vulnerable versions cited. Update your package manager dependencies (npm, yarn, or equivalent) to pull the patched release. Test the upgrade in a non-production environment first, as Content Credentials libraries may interact with other components in your authentication or verification pipeline. Consider automated dependency scanning in your CI/CD to catch outdated versions.

Detection guidance

Monitor system resource utilization (CPU, memory, file descriptors) for sudden spikes or sustained high consumption correlating with requests to services using CAI Content Credentials. Log any resource exhaustion errors or out-of-memory conditions in applications using c2pa-web or c2pa. If you have access to application logs, look for patterns of large or malformed credential processing requests that precede resource exhaustion events. Implement alerting on abnormal process behavior tied to credential verification workflows.

Why prioritize this

Although rated Medium severity, organizations should prioritize patching based on their deployment model. If c2pa-web or c2pa is embedded in a customer-facing or internally critical service, a local attacker gaining system access could halt content verification and authentication workflows. In multi-tenant SaaS environments, the risk escalates because one malicious tenant could impact availability for others. Prioritize patching if the library is in production, exposed to untrusted input, or handles sensitive content authenticity operations.

Risk score, explained

The CVSS 6.2 Medium score reflects high availability impact (complete denial of service) offset by a local attack vector and no privileges required. The vulnerability does not compromise confidentiality or integrity, only availability. In isolated or single-user systems, the risk is lower; in shared or cloud environments where multiple users or applications run on the same infrastructure, the practical risk is higher because a low-privilege attacker can impact other users' services.

Frequently asked questions