← ClaudeAtlas

unit-economics-stress-testlisted

Build-stage skill: models a feature's per-user token cost at three scale points — every number derived from stated assumptions, arithmetic fully reproducible. Use when inference cost needs stress-testing before commitment — 'what does this cost per user at 1k/50k/1M users', 'will inference eat our margin at scale', 'run the unit economics on this feature' — or when /pm routes such a request here. If token counts or prices are missing it asks for them or proposes labeled assumptions — it never silently invents them. Do NOT use for portfolio-level roadmap economics (roadmap-reality-check), pricing-structure design (pricing-tradeoff), bare per-MTok price lookups, or general infra cost reduction.
Abhillashjadhav/PM-agent-OS · ★ 1 · Testing & QA · score 60
Install: claude install-skill Abhillashjadhav/PM-agent-OS
# Unit Economics Stress Test Per-user cost at three scales, every figure re-derivable from the assumptions on the page. If a reader with a calculator can't reproduce a number, it doesn't ship. ## Verification gates (defined first; output is blocked until all pass) - **G1 — Reproducible arithmetic:** every number derives from stated assumptions with the derivation shown (tokens × price / 1M, uses × days, cost × users). Re-running the math from the assumption block must reproduce every figure exactly; rounding is shown, not hidden. - **G2 — No unstated levers:** no bulk discounts, model downgrades, cache rates, or usage-decay curves the input didn't grant. A what-if beyond the stated assumptions appears only as `[ESTIMATE: <assumption> → <arithmetic>]`. - **G3 — Assumptions complete or flagged:** missing load-bearing inputs (token counts, prices, usage rate) are requested or proposed as labeled assumptions the user confirms — never silently defaulted. ## Steps 1. **Freeze the assumption block.** Tokens per use (input/output, with any cacheable share), uses per user per period, contract prices, scale points. This block is the model's ground truth — everything downstream cites it. Anything missing → Step G3 behavior, before any math. 2. **Compute per-use cost:** input tokens × input price / 1M + output tokens × output price / 1M, each term shown. Apply stated cache/batch terms as given, as a separate visible variant — not silently blended. 3. **Compute per-user-per-month:**