BORA Mal dataset: a CentOS 7.9.2009 binary corpus

A 2,702-file binary corpus scraped from a CentOS 7 Google Compute Engine instance, used as the "Mal" dataset. The release is CentOS 7.9.2009, established by hash-matching against vault.centos.org. Contents: 2,440 files byte-identical to CentOS 7.9.2009 os/ and updates/ RPMs, 13 from EPEL 7, 1 installer-written kernel copy, 25 Google-distributed binaries (Cloud CLI, GCE guest environment, bundled CPython 3.9), and 223 artifacts left in the directory by earlier experiments. README.md documents all six categories; manifest/ gives the SHA-256, origin package and upstream URL of every file. Licensing: an aggregation of independently licensed works, not offered under any single licence. The CentOS, EPEL and Google binaries are unmodified redistributions under their upstream licences; all licence texts are in licenses/, and manifest/PACKAGES.tsv cites the corresponding source RPM for every GPL/LGPL package. Documentation and manifests are CC0-1.0. SHA-256 of centos_pack.tgz: 7d7cd1ca57895d3649dfa7320d7b91605772ea771a86954e43ecda4dd2591d4d