animato

This module does not show overt, self-contained malware (no obvious eval/exec/backdoor/obfuscated payloads in the snippet). However, it is explicitly designed to move untrusted Python code—originating from user scripts and/or LLM output—through a local validation step and then submit it to a remote server endpoint (/api/run) for execution. The presence of a documented validation bypass (--skip-gate) and the use of externally supplied/returned code as an execution payload make the supply-chain risk high. Overall risk hinges on the correctness and robustness of validate_bpy_script.py and the server-side enforcement/sandboxing of /api/run and /api/chat.

No direct evidence of explicit malware (e.g., backdoors, hardcoded secrets, or exfiltration endpoints) is present in the provided fragment because it is a specification rather than implementation code. However, it describes a high-impact architecture where LLM-generated bpy scripts are executed via POST /api/run (and server mode executes immediately after model responses). The security posture therefore hinges on the static gate’s completeness and strong runtime sandboxing/allowlisting; additionally, correctness depends on correctly branching on the ok flag since HTTP 200 may be returned on failure. Review the actual /api/run implementation, the static gate, and the Blender execution sandbox to validate safety controls.