Threat database

CVE-2026-108789 is a medium-severity vulnerability in deepseek-ai deepseek-harness versions 0.1.1-rc.0, 0.1.1-rc.1, and 0.1.1-rc.2. It affects the agent function of the dsh-workflow-worker-thread component and allows remote enforcement of behavioral workflow manipulation. Public exploit code exists, but no known exploitation in the wild has been reported. The vendor has not responded to disclosure attempts, and no patch or remediation guidance is currently available.

CVE-2026-108788 is a medium severity vulnerability in deepseek-ai deepseek-harness versions 0.1.1-rc.0, 0.1.1-rc.1, and 0.1.1-rc.2. It involves a sandbox issue in the spawn function of the Sandbox component, triggered by manipulation of the Profile argument. The vulnerability can be exploited locally. The vendor has not responded to disclosure, and no patch is currently available. The exploit has been publicly disclosed but is not known to be used in the wild.

CVE-2026-101102 is a medium severity vulnerability in deepseek-ai deepseek-harness versions 0.1.0-rc.0 through 0.1.0-rc.7. It affects the run_code function in the Code Mode Sandbox component, resulting in a sandbox issue that can be exploited remotely. The vendor explicitly states that the sandbox is for containment only and not a security boundary. There is no vendor response or patch available at this time.

A vulnerability has been found in deepseek-ai deepseek-harness up to 0.1.7-rc.2. Affected is an unknown function of the file packages/sandbox/sandbox-local/src/profiles.ts of the component Landlock Backend. Such manipulation leads to improper isolation or compartmentalization. The attack must be carried out locally. The exploit has been disclosed to the public and may be used. It is advisable to implement a patch to correct this issue. The vendor was contacted early about this disclosure but did not respond in any way.