A simple method was developed for detection and differentiation of five Tier 1 bacterial agents, including Bacillus anthracis, Francisella tularensis, Yersinia pestis, Burkholderia pseudomallei and Burkholderia mallei as well as their closely related near neighbors by gas chromatography-mass spectrometry (GC-MS). Generally, different classes of compounds can be used as biomarkers for biowarfare agent detection, including nucleic acids (i.e., DNA or RNA), proteins (i.e., antibodies), carbohydrates (i.e., sugars), lipopolysaccharides, lipids (i.e., fatty acids) and small molecules. One-step thermochemolysis (TCM) was developed to provide GC-MS detectable biomarker signatures, including sugars, fatty acids and small molecules. Solid phase micro-extraction (SPME) was used for biomarker extraction, concentration and introduction into the GC-MS. Statistical algorithms were constructed using a combination of biomarkers for the five agents, which were robust against different growth conditions (medium and temperature). A general GC-MS temperature program was developed for all five Tier 1 bacteria. The total analysis time, including TCM, SPME extraction and GC-MS, is approximately 40 min. The total-ion chromatograms are very different for the five species. The final goal of this research was to develop an accurate, fast, simple, robust and automated method for field application. Therefore, an automated sample preparation system was designed, constructed and tested. The system automatically controls the movement of sample vials from one position to another, crimping of septum caps onto the vials, precise delivery of reagents and TCM reaction times and temperatures. The specific operations of introduction of sample vials, SPME sampling, injection into the GC-MS system and ejection of used vials from the system were performed manually in this study, although they can be integrated into the automated system. Manual SPME sampling is performed by following visual and audible signal prompts for inserting the fiber into and retracting it from the sampling port. A rotating carousel design allows for simultaneous sample collection, reaction, biomarker extraction and analysis of sequential samples. Bacillus species were used to test this autoreactor, and 96% of the samples were correctly identified using a statistical algorithm. This research applies not only to the rapid identification of Tier 1 agents after a biological attack, but should also benefit clinical diagnosis, which is essential to effective treatment.