Repositories reading strong
Every public repository surveyed, browsable and searchable. Including the unflattering ones — a tool that only published good results would be a badge vending machine, not a diligence instrument.
| Repository | Language | Size | Stars | Read | Band |
|---|---|---|---|---|---|
| gristlabs/grist-core 2 adequate · 3 strong | TypeScript | 561,872 lines | 12,043 | 2026-11-03 | |
| quartznet/quartznet 1 adequate · 4 strong | C# | 608,055 lines | 7,092 | 2026-11-08 | |
| milvus-io/milvus 2 adequate · 3 strong | Go | 2,186,085 lines | 46,451 | 2026-10-22 | |
| rocketchat/rocket.chat 2 adequate · 3 strong | TypeScript | 1,267,556 lines | 46,294 | 2026-10-22 | |
| getagentseal/codeburn 2 adequate · 3 strong | TypeScript | 387,337 lines | 11,670 | 2026-11-03 | |
| genspark-ai/genoffice 2 adequate · 3 strong | TypeScript | 1,395,011 lines | 11,629 | 2026-11-03 | |
| evidence-dev/evidence 2 adequate · 3 strong | TypeScript | 216,191 lines | 7,049 | 2026-11-07 | |
| grafana/pyroscope 1 adequate · 4 strong | Go | 530,836 lines | 11,711 | 2026-11-03 | |
| komari-monitor/komari 2 adequate · 3 strong | Go | 83,108 lines | 6,882 | 2026-11-08 | |
| getsentry/sentry 2 adequate · 3 strong | Python | 2,321,511 lines | 45,628 | 2026-10-22 | |
| certbot/certbot 1 adequate · 3 strong | Python | 104,561 lines | 33,257 | 2026-10-25 | |
| halo-dev/halo 2 adequate · 3 strong | Java | 389,560 lines | 39,970 | 2026-10-23 | |
| nlog/nlog 4 strong | C# | 162,867 lines | 6,548 | 2026-11-08 | |
| glpi-project/glpi 1 adequate · 4 strong | PHP | 1,843,812 lines | 6,560 | 2026-11-08 | |
| centrifugal/centrifugo 1 adequate · 4 strong | Go | 125,526 lines | 10,946 | 2026-11-03 | |
| hazelcast/hazelcast 1 adequate · 3 strong | Java | 2,020,246 lines | 6,629 | 2026-11-07 | |
| triliumnext/trilium 2 adequate · 3 strong | TypeScript | 1,092,422 lines | 38,507 | 2026-10-23 | |
| istio/istio 2 adequate · 3 strong | Go | 850,268 lines | 38,469 | 2026-10-23 | |
| composer/composer 1 adequate · 3 strong | PHP | 154,635 lines | 29,538 | 2026-10-25 | |
| koreader/koreader 2 adequate · 3 strong | Lua | 235,459 lines | 30,020 | 2026-10-25 | |
| firecracker-microvm/firecracker 1 adequate · 4 strong | Rust | 228,438 lines | 37,583 | 2026-10-23 | |
| falkordb/falkordb 1 adequate · 3 strong | Rust | 272,760 lines | 10,502 | 2026-11-03 | |
| rxhanson/rectangle 1 adequate · 3 strong | Swift | 35,697 lines | 29,982 | 2026-10-25 | |
| typeorm/typeorm 2 adequate · 3 strong | TypeScript | 334,851 lines | 36,659 | 2026-10-23 | |
| lightpanda-io/browser 2 adequate · 3 strong | Zig | 384,571 lines | 36,914 | 2026-10-23 | |
| lightdash/lightdash 2 adequate · 3 strong | TypeScript | 2,110,585 lines | 6,164 | 2026-11-08 | |
| pingcap/tidb 2 adequate · 3 strong | Go | 2,035,379 lines | 40,670 | 2026-10-22 | |
| generalaction/emdash 1 adequate · 4 strong | TypeScript | 717,818 lines | 6,171 | 2026-11-08 | |
| redisearch/redisearch 1 adequate · 3 strong | Rust | 653,101 lines | 6,247 | 2026-11-07 | |
| frappe/erpnext 4 strong | Python | 1,013,840 lines | 40,460 | 2026-10-22 |
What a band means. A band sums up one reading of one commit: strong, adequate or significant findings. Here it counts the project's code, process, product, security and operations. A report reads more than that, so its band can differ from its row.
How listing works. Every public repository surveyed is listed — assessing public work is commentary on public work, and a library that hid the small ones would be quietly editorialising about which code counts. Private repositories are never listed. Anyone who proves control of a repo can attach a response to a finding, request a re-run, or delist it. Delisting is immediate and no-questions. An owner's response is removed only when it is unlawful, impersonates someone, publishes personal data or a credential, or is not about the finding. Disagreeing with the report is never a reason. To report a response, or to reach us about one, write to takedown@borehole.dev. There is no org-level score. Repositories are listed one by one, and sorting by band orders them by their own readings. Takedown: one email, 48 hours.