Devolutions Server CVE-2026-9246

Exploitation requires an authenticated session with at least vault read access - unauthenticated exploitation is not possible. The targeted vault entries must be configured as 'sealed', as this is the specific access control mechanism being bypassed. The attacker must craft a specific API request against the entry documentation or attachment endpoint; standard UI interactions do not trigger the bypass. Exploitation is limited to network-accessible Devolutions Server instances; local or air-gapped deployments reduce exposure. SSVC confirms the attack is not automatable, meaning bulk exploitation across many sealed entries would require additional effort per entry.

The official CVSS vector (AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:L/A:N) scores 4.3 Medium but contains an apparent metric inconsistency: the vulnerability description is purely a read/disclosure issue (retrieving documentation and attachments), yet the vector assigns I:L (integrity impact) with C:N (no confidentiality impact). This is likely inverted - the real impact is confidentiality, not integrity. Despite the inconsistency, the score remains 4.3 under either reading. EPSS is 0.03% (8th percentile) and SSVC assigns Exploitation: none and Automatable: no, both confirming minimal real-world exploitation pressure. The threat is realistic only from authenticated insiders or accounts with compromised vault-read credentials - not from external unauthenticated attackers. For PAM environments, however, where sealed entries likely contain high-value credentials or sensitive documentation, the qualitative impact of unauthorized disclosure can substantially exceed what the medium CVSS score implies.

An authenticated employee or contractor with standard vault read access crafts a specific API request targeting the documentation or attachment endpoint of a sealed vault entry they are not authorized to view. The server's access control layer validates the vault-level permission but fails to enforce the additional sealed-entry restriction, returning the protected content. No elevated privileges or special tooling are required beyond knowledge of the API structure and a valid session token.