{
  "name": "FounderMetricBench distribution kit",
  "version": "2026-07-27",
  "status": "prepared_not_published",
  "canonicalUrl": "https://marco-is-my-friend.marcohergee813.chatgpt.site/arena",
  "creator": {
    "name": "Marco Hergi",
    "location": "New York City",
    "profile": "https://marco-is-my-friend.marcohergee813.chatgpt.site/marco-hergi"
  },
  "oneSentenceDescription": "FounderMetricBench is a public 24-case evaluation set inside an 11-tool read-only MCP system for testing whether an AI can route a founder question, calculate the result, retrieve its evidence, and state its safety boundaries.",
  "proofUrls": [
    "https://marco-is-my-friend.marcohergee813.chatgpt.site/arena",
    "https://marco-is-my-friend.marcohergee813.chatgpt.site/founder-metric-bench.json",
    "https://marco-is-my-friend.marcohergee813.chatgpt.site/research/plugin-submission-kit",
    "https://marco-is-my-friend.marcohergee813.chatgpt.site/research/stripe-mcp-benchmark",
    "https://github.com/MARCCHERGGI/indie-metrics-mcp"
  ],
  "youtube": {
    "status": "prepared_not_published",
    "title": "Can an AI Defend a Founder Metric? FounderMetricBench Demo",
    "description": "Marco Hergi runs a public AI evaluation across routing, calculation, retrieval, and safety. Inspect the 24 gold cases, the synthetic evidence, and the 11 read-only MCP tools at the canonical link.",
    "outline": [
      "State the testable question",
      "Show one successful routing and calculation case",
      "Show one failure or boundary case",
      "Open the exact evidence URL",
      "Explain that synthetic data is not financial advice",
      "Invite independent reproduction, not endorsement"
    ]
  },
  "githubRelease": {
    "status": "prepared_not_published",
    "title": "FounderMetricBench public baseline",
    "body": "This release publishes 24 gold cases across routing, calculation, retrieval, and safety, plus stable evidence URLs and a public synthetic MCP endpoint. Results and known gaps remain visible. No real Stripe account, credentials, or customer data are used."
  },
  "directoryRecord": {
    "status": "prepared_not_submitted",
    "category": "AI evaluation and developer tooling",
    "description": "A reproducible founder-metrics benchmark with transparent formulas, synthetic data, stable citations, and a read-only Streamable HTTP MCP endpoint."
  },
  "outreach": {
    "status": "prepared_not_sent",
    "subject": "A reproducible AI benchmark for founder metrics",
    "message": "I built FounderMetricBench, a public 24-case evaluation for routing, calculation, retrieval, and safety in founder-metric questions. Every claim links to a stable source or synthetic fixture. If it is useful to your research, you can reproduce the cases and report failures independently."
  },
  "prohibitedClaims": [
    "Do not claim that OpenAI, Anthropic, Perplexity, Google, Stripe, or any directory endorses the project.",
    "Do not claim guaranteed indexing, ranking, citations, referrals, traffic, funding, or founder interest.",
    "Do not describe synthetic results as real business performance.",
    "Do not describe prepared materials as published, submitted, or sent."
  ]
}