Update to MayRetire

{
  "description": "Compare, optimize, stress-test, simulate, or backtest an existing MayRetire plan through the remote MCP when the user asks what-if or risk questions.",
  "included_files": [
    {
      "relative_path": "references/scenario-design.md",
      "size_in_bytes": 2069
    },
    {
      "relative_path": "references/stress-and-simulation.md",
      "size_in_bytes": 2300
    }
  ],
  "name": "mayretire-test-scenarios",
  "skill_md_contents": "---\nname: mayretire-test-scenarios\ndescription: Compare, optimize, stress-test, simulate, or backtest an existing MayRetire plan through the remote MCP when the user asks what-if or risk questions.\n---\n\n# Test MayRetire scenarios\n\nThe remote service needs the complete user-supplied plan on each plan call;\nit cannot fetch the user's MayRetire account or browser plan. Explain that\ntool calls send the plan through the assistant to MayRetire for calculation.\nPreserve the original plan, income and estate requirements, and all excluded\nvariables. Ask which consequential areas may change when the request leaves\nthat unclear. Inspect and validate before relying on results.\n\nRead [scenario design](references/scenario-design.md) before edits or search.\nConsult the bundled\n[MayRetire tutorial](../mayretire-app-assistant/references/MayRetire_Tutorial.md)\nwhen the user's question needs an explanation of a scenario control or screen.\nUse `describe` and the inspected plan for JSON field meaning and active\nsettings; use calculated results for outcomes. Do not treat a general tutorial\nexample as evidence about this user's plan.\nUse `compare_plan_payloads` for small independent patches and deterministic\nscreening. Use `edit_plan_payload` for a selected complete plan or item\noperations; show the diff and validate the result. For larger exploration or\noptimization, use `start_exploration_job` or `start_optimization_job` with a bounded request, poll\n`status_plan_job`, collect `result_plan_job`, and retrieve only selected\ncandidate plans with `get_candidate_plan`. Cancel work the user no longer\nneeds. Search results describe the evaluated candidate space, not every\npossible plan.\n\nRead [stress, Monte Carlo, and backtesting](references/stress-and-simulation.md)\nbefore a risk test. Run the requested test rather than substituting a different\nmetric. Distinguish six-scenario stress pass counts and average spending funded\nfrom Monte Carlo probability. Compare plans with aligned success criteria,\ntrial count, seed, and engine build. Use only returned metrics and explain\nsampling uncertainty. An apparent improvement may come from reduced spending;\nshow that tradeoff. Do not give personalized financial or tax advice.\n"
}