A flaw affecting FoundationAgents MetaGPT up to version 0.8.1 allows code injection through the generate_thoughts function in tot.py. The description indicates that manipulation of this function leads to code injection, which can be triggered remotely. Based on the description, it is inferred that an attacker can craft malicious input that is evaluated by the system, resulting in the execution of arbitrary code and full compromise of the host running MetaGPT.

All installations of FoundationAgents MetaGPT versions ≤0.8.1 contain the vulnerable code in the Tree‑of‑Thought Solver component. The flaw resides in tot.py, and any deployment that imports or uses generate_thoughts is potentially exposed. No patch or mitigation is available from the vendor at this time, so the issue remains active for all affected releases.

The CVSS score of 6.9 denotes moderate severity, yet the vulnerability permits remote code execution. The exploit code is publicly available and may be employed by attackers. Although an EPSS score is not published, the combination of remote accessibility and lack of vendor response raises the likelihood of exploitation. It is inferred that the attack vector is remote exploitation of services that expose the vulnerable function. The vulnerability is not yet listed in the CISA KEV catalog, but the potential for critical impact warrants high priority action.