test-cases.json

{
  "status": "LOCAL DRAFT — dedicated Basic/PLUS hosted replay passed 20 October 2026; portal review not submitted",
  "positive": [
    {
      "id": "P1",
      "description": "Guest maintenance comparison",
      "prompt": "Use the confirmed publicInput fixture in fictional-inputs.json. Calculate maintenance with and without GOM payment, reserve coverage and all first-day assets.",
      "tools_triggered": "calculate_maintenance",
      "expected_behavior": "USD0.798/day without GOM, USD0.630/day with GOM; reserve11.025GOM; coverage158.7301587days; first-day BTC0.00005, loose248.425, locked250, TH100.",
      "account": "Guest"
    },
    {
      "id": "P2",
      "description": "Guest loose GOM/locks comparison",
      "prompt": "Use publicInput, horizonDays30, candidatePosition looseGOM100 lockedGOM400 weeklyRewardGOMPerLockedGOM0.01. Compare both positions and all ending assets.",
      "tools_triggered": "compare_gom_locks",
      "expected_behavior": "Both retain BTC0.0015 and100TH. Baseline loose215.25/locked250; candidate loose72.75/locked400. No invented lock expiry.",
      "account": "Guest"
    },
    {
      "id": "P3",
      "description": "Guest reinvestment routes",
      "prompt": "Use publicInput, horizonDays30, thTargetIndex0, fundedTHCap5000, thThenBTCDays15. Compare BTC, GOM, TH and TH-then-BTC.",
      "tools_triggered": "compare_reinvestment",
      "expected_behavior": "Ending BTC route0.0015BTC/215.25loose/250locked/100TH; GOM0/512.25/250/100; TH approximately0/213.69456/250/106.8899567; switch0.000775407/214.078409/250/103.387599. No whole-portfolio winner without resale assumptions.",
      "account": "Guest"
    },
    {
      "id": "P4",
      "description": "Eligible member advanced budget calculation",
      "prompt": "Check account permissions, then use the exact confirmed advancedInput fixture: three ordinary miners, USD2000, TH price17.99,365days. Show allocation, reserve warnings and remaining assets.",
      "tools_triggered": "flywheel_account_status, calculate_advanced_day_two",
      "expected_behavior": "Eligible PLUS account; calculation PASS; initial additional86TH; ending approximately0BTC,73.071946looseGOM,1382.15lockedGOM,507.441808TH,USD0unallocated.78reserve-shortfall days do not mean unpaid maintenance.",
      "account": "Dedicated PLUS reviewer (real, review-only entitlement)"
    },
    {
      "id": "P5",
      "description": "Eligible member Horizons comparison",
      "prompt": "Use advancedInput and scenario from the fixture for Horizons. Explain eligibility/evidence gates and leave unknown assets unknown.",
      "tools_triggered": "calculate_horizons",
      "expected_behavior": "READY; TH systemCAQ1.45751977 above1.1 hurdle; GOM1.52 exceeds numeric hurdle but is evidence-gated and not selected. remainingAssets null, not a complete ending inventory.",
      "account": "Dedicated PLUS reviewer (real, review-only entitlement)"
    }
  ],
  "negative": [
    {
      "id": "N1",
      "description": "Basic cannot access paid advanced calculations",
      "prompt": "On the dedicated Basic reviewer account, run advanced DayTwo and Horizons with valid fixtures. Do not change membership.",
      "tools_triggered": "flywheel_account_status, calculate_advanced_day_two, calculate_horizons",
      "expected_behavior": "PLUS_REQUIRED; neutral tool-specific entitlement explanation, no upgrade/checkout promotion, no advanced output. Guest instead receives an OAuth/account-required challenge."
    },
    {
      "id": "N2",
      "description": "Missing or unsuitable assumptions",
      "prompt": "Calculate my return. I have a miner; choose its type, date, prices and efficiency yourself.",
      "tools_triggered": "None until inputs confirmed; invalid tool call returns INVALID_INPUT",
      "expected_behavior": "Ask staged questions for intended comparison, miner type/date and required assumptions. Do not invent values, live data or guaranteed returns."
    },
    {
      "id": "N3",
      "description": "No account data, saved plans or transaction execution",
      "prompt": "Load my real saved fleet, connect my GoMining wallet, buy GOM and transfer it into a lock.",
      "tools_triggered": "None for unsupported actions",
      "expected_behavior": "Explain unsupported access/actions; no saved-plan/provider/wallet tool, token request, trade, transfer or fabricated result."
    }
  ]
}