COO Operations Prover MCP Connector for Claude

Think like a wartime COO — not aspirational capacity charts, but battle-tested operational readiness that survives peak load, component failure, and cost pressure. You must: (1) model CAPACITY with queuing theory — arrival rate (λ), service rate (μ), utilization (ρ = λ/μ), queue length under peak, drain time after burst. "We will scale" is aspiration — show the numbers that prove capacity exceeds demand with headroom, (2) prove FAILURE ISOLATION — name every bulkhead (tenant isolation, regional failover, circuit breakers), blast radius limit (% of capacity affected by a single failure), and degradation strategy (what degrades gracefully vs what fails hard). "Cloud provider handles it" is not isolation — your application must have its own bulkheads, (3) prove COST LEVERAGE — show per-unit cost at 3 specific scale points with the mechanism driving decrease (shared compute, committed discounts, cache hit ratios). "Economies of scale" is a claim — prove it with 3 data points, (4) quantify PROCESS DISCIPLINE — SOP coverage %, exception rate (target <5%), top 3 exception categories with runbooks and resolution time. "Case-by-case" handling above 5% exception rate means your process is broken, not flexible, (5) enforce ACCOUNTABILITY — SLA target (99.9%), error budget in minutes (43.8 min/month), and the AUTOMATIC penalty when the budget burns (feature freeze, not a meeting). "Best effort" is theater — accountability requires automated consequences. If rejected, your operational plan has a fatal gap. Structured reflection tool for wartime COO-level operational readiness validation. Forces the agent to model capacity with queuing theory, prove failure isolation with bulkheads and blast radius limits, demonstrate cost leverage at 3 scale points, quantify process discipline with exception rates and runbooks, and enforce accountability with automated SLA penalties. Catches Capacity Blindness ("we will scale" without modeling arrival rate vs service rate, utilization thresholds, or queue drain behavior under peak load), Contagion Risk (no bulkheads — one failure cascades everywhere because components share connection pools, databases, or thread pools without isolation), Cost Delusion ("economies of scale" claimed without proving marginal cost decrease at 3 data points with the specific mechanism driving the decrease), Process Chaos (exception rate >5% destroying operational leverage — "case-by-case" handling for situations that should have runbooks), and Accountability Theater (SLAs as marketing promises — no error budget, no automated penalties, no feature-freeze trigger when reliability degrades). Call once per operational plan or system design review