This repository is a curated data product, not a software application: it contains zero executable code. There are no scripts, no CI/CD pipeline, no bot source, and no tests — only 12 flat-text data files and supporting metadata. The README states that a bot automatically probes public trackers daily and commits fresh lists, but that bot lives outside this repository; only its outputs are versioned here. The repo is essentially a continuously refreshed dataset of public BitTorrent tracker announce URLs, consumed directly by torrent clients via raw GitHub, GitHub Pages, and jsDelivr CDN mirrors. The data lineage is visible only through the output artifacts. Input is a community-maintained pool of public trackers (sourced from issue submissions and sites like newtrackon). The external bot performs health checks, latency measurement, protocol classification (UDP/HTTP/HTTPS/WebSocket/I2P/Yggdrasil), deduplication by domain and resolved IP, DNS resolution into IP-only variants, Cloudflare IP filtering, and removal of trackers flagged as malfunctioning, fake-seeding, antivirus-detected, private/registered-torrents-only, deprecated, or duplicate. The outputs are protocol-split lists (trackers_all_udp.txt, trackers_all_http.txt, etc.), a quality-ranked top-20 list (trackers_best.txt), a full list (trackers_all.txt), and IP-resolved variants for DNS-avoidance scenarios. The curation quality is genuinely high: the blacklist.txt encodes removal decisions with categorized, human-readable reasons, and the dual-domain/IP output pattern solves a real operational problem for torrent clients. However, because the pipeline code is absent, an analyst cannot reproduce, extend, or re-run the generation process — the repository is a consumable data artifact, not a research pipeline or toolkit. It is not a hybrid repo; it is purely a data product with no application component.