CVE Vulnerability Catalog

FFmpeg through version 8.1.2, fixed in commit b506faf, contains a heap out-of-bounds write vulnerability in the native PNG and APNG encoders. An attacker can supply a crafted PNG image with a malicious eXIf chunk, causing tens of thousands of bytes to be written past the buffer boundary, leading to heap corruption and potential code execution.

FFmpeg through version 8.1.2, fixed in commit aafb5c6, contains a signed integer overflow vulnerability in the MACE6 audio decoder. An attacker can supply a crafted CAF file with a malicious bytes_per_packet value, causing an integer overflow and undersized buffer allocation, leading to a heap out-of-bounds write and potential code execution.

In FFmpeg up to version 8.1.2, there is a heap out-of-bounds write vulnerability in the vf_hqdn3d filter. An attacker can supply a crafted video where frame resolution increases between frames and filter reinitialization is disabled, leading to heap memory corruption.

In epa4all prior to version 2026-05-20, an attacker on the network path can present a self-signed TLS certificate and intercept the connection. For non-VAU connections, this allows direct read and modification of inner traffic, including smartcard operations and OIDC authentication exchanges.

In Hulumi prior to version 1.4.0, drift detection in CI/cron could silently cache transient adapter failures as 'all clear', masking real attacks for up to six hours, or falsely escalate ordinary provider-version churn to incident severity.

In Hulumi prior to version 1.4.0, users of AccountFoundation could deploy an AWS account whose CloudTrail/Config audit logs were deletable by any S3-delete-capable principal, despite the startup-hardened tier guaranteeing tamper-resistance. Sandbox deployments had no audit immutability.

In Hulumi prior to version 1.4.0, there is a bypass via decoy sibling resources targeting a different bucket.

In Hulumi prior to version 1.4.0, policy packs can be bypassed by a forged Pulumi-URN logical name.

In Hulumi prior to version 1.4.0, IAM-role policy checks can be bypassed when the role trusts multiple OIDC providers.

In the service-proxy component of Red Hat Advanced Cluster Management for Kubernetes (RHACM) and multicluster-engine (MCE), the proxy appends impersonation headers without removing caller-supplied values. An authenticated hub principal can inject an Impersonate-Group header to escalate to cluster-admin on every managed cluster.

Pre-authentication heap buffer overflow in libssh2 up to 1.11.1, allowing a malicious SSH server to corrupt client heap metadata by sending a packet with packet_length smaller than cipher block size during Encrypt-then-MAC negotiation.

Missing bounds check in libssh2 up to 1.11.1, allowing a malicious SSH server to trigger an arbitrary-length heap out-of-bounds read and free of an uninitialized pointer via the publickey subsystem.

Pre-authentication integer underflow in ssh2_cipher_crypt() in libssh2 up to 1.11.1, allowing a malicious SSH server to crash any connecting client by negotiating AES-GCM ciphers.

Double-free vulnerability in sftp_open() in libssh2 up to 1.11.1, allowing a malicious SSH server to corrupt the heap of any authenticated client opening an SFTP session.

OS command injection vulnerability in sysPass up to version 3.2.11, allowing authenticated administrators to execute arbitrary commands by setting a malicious backup path and triggering a backup.

Missing authorization vulnerability in sysPass up to version 3.2.11, allowing authenticated users with the PUBLICLINK_CREATE flag to trigger unauthorized decryption and persistent storage of any vault account's password.

Missing object-level authorization in the JSON-RPC API in sysPass up to version 3.2.11, allowing API token holders to enumerate account metadata, overwrite passwords, and delete accounts across the entire vault without per-account access control.

Insecure direct object reference vulnerability in sysPass up to version 3.2.11, allowing any authenticated attacker to access account file attachments belonging to accounts they do not have ACL permissions for.

Inefficient algorithmic complexity vulnerability in the bandit library (Elixir), allowing unauthenticated remote denial of service via CPU exhaustion during WebSocket fragment reassembly.

Suna before version 0.9.102 contains a broken access control vulnerability in the message queue API. Authenticated attackers can access and manipulate queue resources belonging to other users by exploiting missing ownership and account isolation checks.