This vulnerability in FoundationAgents MetaGPT up to version 0.8.2 involves code injection via the DataInterpreter function in metagpt/roles/di/data_interpreter.py. Exploitation requires local access with low privileges and no user interaction. The vulnerability has been publicly disclosed, but the vendor has not issued a fix or advisory. The CVSS 4.0 vector indicates low attack complexity and privileges required, with partial impact on confidentiality, integrity, and availability.

Successful exploitation allows an attacker with local access to inject and execute arbitrary code within the MetaGPT environment. This could lead to unauthorized code execution and potential compromise of the affected system. The impact is limited by the requirement for local access and low privileges, reducing the likelihood of remote exploitation.

No official patch or remediation is currently available from the vendor. Since the vulnerability requires local access, restricting local user permissions and access to trusted users only can reduce risk. Monitor for any future vendor advisories or patches and apply them promptly once available.