Answers are grounded in this paper and cite their sources.
Developer recruitment attacks hide executable commands in verification pages, assessment repositories and editor extensions. Static inspection is difficult when clipboard handlers assemble commands through helper functions and obfuscators replace configuration strings with calls into encoded, rotated arrays. We built 'unk-scanner' to recover these structures without executing the submitted code. It follows clipboard values through the abstract syntax tree and evaluates a restricted subset of expressions. For recognised obfuscator loops, it recovers array order by reproducing the checksum condition in a restricted evaluator, with explicit assumptions governing the correctness of the recovered rotation. Detection rules combine indicators to distinguish suspicious execution paths from techniques also used by legitimate developer tools. The evaluation covers approximately 2.8 million files from software registries and a public malware dataset, retained UNK_DeadDrop artifacts, controlled obfuscation experiments and evasion fixtures. The isolated scanner processed 272 calibration files per second. Two error-level alerts among 19,976 recently changed npm packages were false positives. Controlled experiments across five obfuscator releases reached 99.7 to 99.8 percent string recovery with the revised engine; every decoded call site matched its original value. Detection coverage remained limited: that engine warned on 128 of 2,122 public malware samples. Small source rewrites also exposed missed behaviours and an incorrect rotation assumption. Recovered strings made hidden code inspectable. Accurate recovery nevertheless coexisted with low malware coverage and easily evaded rules.
No takes yet. Share an insight, caveat, or question.
Ilja Krasnjanski (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: