Back Of The Envelope
Turns "high traffic" into QPS, storage, bandwidth and server-count numbers, with a script to check the math.
Getest · Werkt
Wat het doet
Guides QPS, storage, bandwidth and server-count estimation for system design using standard rule-of-thumb tables (per-server QPS, latency numbers, powers of two). Triggers on estimate/capacity-planning language, and even implicitly whenever a design choice hinges on scale. Ships a deterministic Python calculator (botec.py) with a golden-fixture test.
Testrapport
Ran the bundled botec.py on a 50M-DAU photo-app scenario: it produced peak QPS, TB/day storage, PB total, and 1.2GB/s bandwidth with labeled peak factor and units, where an unaided baseline only got avg QPS and total storage — genuinely more complete. Own test_botec.py golden-fixture check passed live. Minor flaw: server-count rounds to "0" instead of "1" for light workloads.
Getest op: 2026-07-15 · Claude Code 2.x (agent harness)
Installatie
git clone https://github.com/proyecto26/system-design-skills mkdir -p ~/.claude/skills cp -r system-design-skills/skills/back-of-the-envelope ~/.claude/skills/back-of-the-envelope
Commando's en voorbeeldprompts
/back-of-the-envelopeTurns "high traffic" into QPS, storage, bandwidth and server-count numbers, with a script to check the math.
Skills reageren op gewone verzoeken — geen commando's om te onthouden. Na installatie activeren prompts zoals deze de skill (in het Engels):
estimate the QPS and storage needs for 50 million daily usershow many servers would I need to handle this traffic loadcalculate bandwidth and storage for a photo-sharing app at scale