Engineering Attack Vectors and Detecting Anomalies in Additive Manufacturing

Engineering Attack Vectors and Detecting Anomalies in Additive Manufacturing⋆ Md Mahbub Hasan1[0009−0002−2226−6494], Marcus Sternhagen1[0009−0001−5605−6631], and Krishna Chandra Roy1[0000−0001−9388−8042] New Mexico Institute of Mining and Technology {mdmahbub.hasan,marcus.sternhagen}@student.nmt.edu, krishna.roy@nmt.edu Abstract. Additive manufacturing (AM) is rapidly integrating into crit- ical sectors such as aerospace, automotive, and healthcare. However, this cyber-physical convergence introduces new attack surfaces, especially at the interface between computer-aided design (CAD) and machine exe- cution layers. In this work, we investigate targeted cyberattacks on two widely used fused deposition modeling (FDM) systems, Creality’s flag- ship model K1 Max, and Ender 3. Our threat model is a multi-layered Man-in-the-Middle (MitM) intrusion, where the adversary intercepts and manipulates G-code files during upload from the user interface to the printer firmware. The MitM intrusion chain enables several stealthy sab- otage scenarios. These attacks remain undetectable by conventional slicer software or runtime interfaces, resulting in structurally defective yet ex- ternally plausible printed parts. To counter these stealthy threats, we propose an unsupervised Intrusion Detection System (IDS) that ana- lyzes structured machine logs generated during live printing. Our defense mechanism uses a frozen Transformer-based encoder (a BERT variant) to extract semantic representations of system behavior, followed by a contrastively trained projection head that learns anomaly-sensitive em- beddings. Later, a clustering-based approach and a self-attention autoen- coder are used for classification. Experimental results demonstrate that our approach effectively distinguishes between benign and compromised executions. Keywords: Cyber-Physical Systems · Intrusion Detection System (IDS) · 3D Printer Attacks · Additive Manufacturing · Anomaly Detection. 1 Introduction Additive Manufacturing (AM), commonly known as 3D printing, has revolu- tionized modern manufacturing by enabling the rapid prototyping and produc- tion of complex components with minimal material waste [1, 2]. Its integration into safety-critical domains such as aerospace, healthcare, automotive, and de- fense has made the underlying infrastructure of AM systems a prime target ⋆Supported by NSF - National Science Foundation arXiv:2601.00384v1 [cs.CR] 1 Jan 2026 2 Hasan et al. for cyber-physical threats [3,4]. With the proliferation of network-connected 3D printers [5], adversaries can exploit attack vectors across the digital thread from CAD design, STL/G-code translation, network communication, to firmware exe- cution, potentially sabotaging the mechanical integrity, dimensional accuracy, or intellectual property (IP) [6] of printed parts [7,8]. Recent studies have demon- strated the feasibility and consequences of attacks at various stages of the AM pipeline. These include manipulations of CAD/STL files [9], malicious firmware modifications [10, 11], G-code-level sabotage [12], and side-channel IP leakage through power, acoustic, and magnetic emissions [6, 8, 13, 14]. Meanwhile, at- tack detection frameworks have been proposed using process monitoring [15,16], statistical modeling [7], and analog emission analysis [4]. Despite these efforts, most prior approaches [17] rely on either predefined attacker models or static analysis or require access to a golden reference model(STL).Hence, a significant gap persists in detecting real-time, stealthy attacks that operate directly at the G-code level, the machine-readable instruction. G-code is generated by slicer software from a CAD model and executed during fabrication. Once uploaded, the assumption of trust in the G-code pipeline creates a critical vulnerability. Unlike STL manipulations, G-code-level attacks can subtly alter toolpaths, ex- trusion volumes, or print timing without triggering visual inspection or violating basic geometry constraints. In this work, we explore stealthy G-code manipulation strategies that by- pass traditional STL-based validation and compromise the final print without overt disruption. We identify three strategies that exploit realistic threat models in networked 3D printing setups, Deferred Print Exploit, Access-Jammed G-code Swap, and Execution-Phase Tampering. Each of these approaches was implemented in a realistic threat model, such as Under-extrusion, Over- extrusion, Noisy G-code Injection, Dimensionality Change, and Internal Cavity Insertion. Assuming the adversary has access to the printer’s file system or con- trol interface (e.g., compromised print servers, remote access tools, or insider threats). We conducted experiments on two widely used FDM platforms, Cre- ality K1 Max and Creality Ender 3, with distinct system architectures (Klipper and OctoPrint, respectively), demonstrating the feasibility and cross-platform generalizability of these threats. A central challenge li

…(Full text truncated)…

This content is AI-processed based on ArXiv data.