CVE Alert: CVE-2026-117568 – bitscorp-mcp – mcp

mcp-ffmpeg through 1.0.5 contains an OS command injection vulnerability that allows attackers to run shell commands by supplying crafted videoPath or outputDir values to the resize-video, extract-audio and get-video-info tools. Attackers can embed $(), backticks or a closing double quote in an existing file or directory name to break out of the child_process.exec command string and execute commands as the MCP server user.

**Risk verdict:** High-impact command execution risk, but urgency cannot be fully ranked because KEV, SSVC, PoC and EPSS status were not provided.

**Why this matters:** Successful exploitation could let an attacker run commands with the service account’s permissions, potentially exposing media, credentials or other accessible data, or disrupting processing. The practical impact depends on what that account can access and whether the service handles inputs from untrusted users or agents.

**Most likely attack path:** An attacker needs local access or a way to submit tool inputs, plus a suitably crafted existing file or directory name; exploitation is otherwise constrained by high complexity and a stated attack requirement. No user interaction is required once the tool processes the input, and the assessed scope is unchanged, so expect compromise within the server’s security context rather than automatic cross-boundary impact.

**Who is most exposed:** Prioritise MCP deployments that accept requests from multiple users or agents, especially shared automation hosts with broad filesystem access.

Alert on shell metacharacters or unexpected quotes in tool path arguments.

Review child-process execution and shell logs for unusual commands from the service account.

Check for unexpected file changes, outbound connections or access to secrets after tool use.

Mitigation and prioritisation:

Obtain KEV, SSVC, PoC and EPSS status before final ranking; treat as priority 1 if KEV is true or EPSS is at least 0.5.

Apply a vendor-confirmed fixed release promptly; confirm remediation availability before scheduling.

Until patched, disable affected tools or restrict them to trusted callers and validated paths.

Run the service with minimal filesystem access and privileges; test changes in staging before rollout.

A considerable amount of time and effort goes into maintaining this website, creating backend automation and creating new features and content for you to make actionable intelligence decisions. Everyone that supports the site helps enable new functionality.

If you like the site, please support us on Patreon or Buy Me A Coffee using the buttons below.