CVE-2026-4784

A vulnerability in the email parsing of Cisco AsyncOS Software for Cisco Secure Email Gateway is due to insufficient validation in the parsing logic. An unauthenticated remote attacker could send a crafted email containing malicious SQL statements, potentially leading to execution of arbitrary SQL statements and command execution with root privileges on the underlying operating system.

The vulnerabilities tracked by CVE-2026-76443 relate to improper neutralization issues grouped under CWE Pillar CWE-707 and were found during Cisco's internal security review. They affect Cisco Secure Email Gateway and Cisco Secure Email and Web Manager.

The vulnerabilities tracked by CVE-2026-76441 relate to improper access control issues grouped under CWE Pillar CWE-284 and were found during Cisco's internal security review. They affect Cisco Secure Email Gateway and Cisco Secure Email and Web Manager.

The vulnerabilities tracked by CVE-2026-76440 relate to path traversal issues grouped under CWE Pillar CWE-23 and were found during Cisco's internal security review. They affect Cisco Secure Email Gateway and Cisco Secure Email and Web Manager.

The vulnerabilities tracked by CVE-2026-20353 relate to improper control of a resource through its lifetime grouped under CWE Pillar CWE-664 and were found during Cisco's internal security review. They affect Cisco Secure Email Gateway and Cisco Secure Email and Web Manager.

In the Yayson library for serializing and reading JSON API data in JavaScript, prior to 4.3.0, Store and LegacyStore use attacker-controlled JSON:API type, id, and relationship names as keys in plain-object lookup tables. A document whose type is __proto__ causes model-cache writes to modify Object.prototype, with the attacker controlling the polluted property name through id and its value through attributes. This can result in process-wide prototype pollution, causing denial of service and logic corruption.

In PraisonAI, a multi-agent teams system, prior to 4.6.62, src/praisonai/praisonai/tools/multiedit.py passes the LLM-controlled filepath parameter directly to open for reading and writing without traversal rejection, symlink resolution, a workspace boundary, or protected-path checks. Prompt-influenced agents can read or overwrite files.

In PraisonAI, a multi-agent teams system, prior to 4.6.58, praisonai.jobs.server.create_app mounts praisonai.jobs.router.create_router under /api/v1/runs without authentication or per-job authorization. Network clients can submit attacker-controlled prompts and agent configuration, list and read jobs, stream results, and cancel or delete other jobs.

The LdapAuth and LinOTPAuth plugins in MISP contain an authentication bypass vulnerability. Missing input validation allows empty or non-string credentials to reach authentication logic, leading to unauthenticated LDAP binds (RFC 4513) or acceptance of empty passwords. This can allow an attacker to impersonate any known user without a password.

Bifrost registers MCP clients through its management API, where a stdio client is a command plus args that is started immediately upon client addition, with no MCP handshake required. Authentication is disabled by default (governance.auth_config.is_enabled=false), meaning every caller is treated as a local admin. A single unauthenticated POST /api/mcp/client is enough to run a program as the Bifrost process user (appuser on the official image).