{"id":"W4400065514","doi":"10.1016/j.cose.2024.103971","title":"Detecting command injection vulnerabilities in Linux-based embedded firmware with LLM-based taint analysis of library functions","year":2024,"lang":"en","type":"article","venue":"Computers & Security","topic":"Security and Verification in Computing","field":"Computer Science","cited_by":15,"is_retracted":false,"has_abstract":false,"ca_institutions":"Carleton University","funders":"","keywords":"Firmware; Computer science; Taint checking; Operating system; Embedded system; Software","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0005541398,0.0002582415,0.0004368769,0.001248974,0.000234089,0.0004341535,0.0006495736,0.000119826,0.0000314],"category_scores_gemma":[0.00004758183,0.0002428759,0.0002389772,0.004594671,0.0001343775,0.0006824512,0.0001688512,0.0004929471,0.000004088874],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001196481,"about_ca_system_score_gemma":0.0003268826,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001788408,"about_ca_topic_score_gemma":0.0002409073,"domain_scores_codex":[0.9977193,0.0003387717,0.0005622767,0.0007169079,0.000342921,0.0003197969],"domain_scores_gemma":[0.9976262,0.001241365,0.0001671868,0.0007493109,0.0001174328,0.00009847385],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0001110149,0.0006767782,0.0255819,0.0007709953,0.0005671656,0.00005879674,0.01352042,0.926952,0.00005442445,0.0111528,0.001239844,0.01931385],"study_design_scores_gemma":[0.0003626939,0.0002790779,0.005561073,0.0002682577,0.00009631995,0.000004882478,0.0003609391,0.9904602,0.0008645675,0.0007895502,0.0006799907,0.0002724746],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.3469951,0.0002581233,0.6509009,0.0005327797,0.0003896097,0.0001949565,0.00001527327,0.00054495,0.0001683327],"genre_scores_gemma":[0.9818613,0.000002071736,0.01767798,0.0002596439,0.00008462745,0.00002785789,0.00006023751,0.00001749379,0.000008793381],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.6348662,"threshold_uncertainty_score":0.9904191,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01266985832429048,"score_gpt":0.2408840868425788,"score_spread":0.2282142285182883,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}