plugin.json

{
  "$schema": "https://agent-plugins.org/schemas/1.0.0/plugin.schema.json",
  "name": "layerz",
  "version": "1.0.2",
  "description": "Build, edit and audit structured financial models (P&L, cash flow, balance sheet) from ChatGPT or Codex, then export them to Excel with live formulas.",
  "author": {
    "name": "Layerz",
    "email": "contact@layerz.cc",
    "url": "https://layerz.cc"
  },
  "homepage": "https://layerz.cc/for-agents",
  "license": "Proprietary",
  "keywords": [
    "finance",
    "financial model",
    "business plan",
    "forecast",
    "fp&a",
    "excel",
    "spreadsheet"
  ],
  "extensions": {
    "com.openai": {
      "interface": {
        "displayName": "Layerz",
        "shortDescription": "Financial models that persist",
        "longDescription": "Layerz is a structured spreadsheet for financial models, built to be driven by an AI assistant. Every line of a model is a named variable with its formula, its timeline and its dependencies, so the assistant can change an assumption and every downstream figure recomputes, instead of rewriting a sheet by hand.\n\nWith Layerz in ChatGPT or Codex you can:\n- List and read your models: structure, formulas, dependencies and computed values for any period.\n- Build a new model from a description or from a template (SaaS, business plan, three statements).\n- Update assumptions and formulas, and see the impact on revenue, EBITDA, cash or any output.\n- Work on scenarios as branches, compare them, and review the version history of each line.\n- Import actuals from an Excel or CSV file and map them to the model.\n- Export a clean .xlsx with live formulas, or share a model with a collaborator.\n\nEach model carries a FINANCE.md file (currency, units, sign conventions, statement structure) that the assistant reads first, so the context persists from one conversation to the next.\n\nWho it is for: finance professionals, founders and advisors who build business plans, budgets, forecasts and deal models.\n\nLimitations: Layerz does not move money, place trades or give investment advice. It only reads and edits the models of the signed-in Layerz account. Bank and accounting connections, when the user sets them up in Layerz, are read-only imports of actuals.",
        "developerName": "Layerz",
        "category": "Finance",
        "capabilities": [
          "Read financial models",
          "Edit assumptions and formulas",
          "Build models from templates",
          "Compare scenarios",
          "Import actuals",
          "Export to Excel"
        ],
        "websiteURL": "https://layerz.cc",
        "supportURL": "https://layerz.cc/for-agents",
        "privacyPolicyURL": "https://layerz.cc/privacy",
        "termsOfServiceURL": "https://layerz.cc/terms",
        "defaultPrompt": [
          "List my Layerz models and summarise the P&L of the latest one.",
          "Build a 5-year SaaS business plan in Layerz from my assumptions.",
          "Raise churn to 2.5% from March and show the impact on year-end cash."
        ],
        "brandColor": "#0A0A0A",
        "brandColorDark": "#C9A74E",
        "composerIcon": "./assets/icon.png",
        "logo": "./assets/logo.png"
      },
      "review": {
        "test_cases": {
          "positive": [
            {
              "description": "List the models available to the signed-in account.",
              "prompt": "List my Layerz models.",
              "tools_triggered": "layerz_list_models",
              "expected_behavior": "Returns the three demo models of the test account: Acme SaaS - Reporting & Forecast, Coffee Shop NYC - Business Plan, Omni Brand - Channel Mix."
            },
            {
              "description": "Read a model and explain its structure and computed values.",
              "prompt": "Open the Coffee Shop NYC business plan in Layerz and summarise its P&L for each year: revenue, main costs and net income.",
              "tools_triggered": "layerz_list_models, layerz_read",
              "expected_behavior": "Reads the model and summarises the P&L by year with figures taken from the model's computed values, explaining that revenue is driven by Customers per Day, Average Ticket and Operating Days per Year."
            },
            {
              "description": "Change an assumption and report the recomputed impact.",
              "prompt": "In the Coffee Shop NYC model, raise the Average Ticket by 10% in every year and tell me how net income and closing cash change.",
              "tools_triggered": "layerz_read, layerz_patch",
              "expected_behavior": "Updates the Average Ticket assumption, then reports the before and after values of net income and closing cash as recomputed by Layerz. Only that assumption is modified."
            },
            {
              "description": "Explain what drives an output through the dependency graph.",
              "prompt": "In the Acme SaaS model, what does the Monthly Churn Rate affect? List the lines that depend on it.",
              "tools_triggered": "layerz_read, layerz_dependencies",
              "expected_behavior": "Lists the downstream lines that depend on Monthly Churn Rate (recurring revenue, retention metrics and the statements fed by them), based on the model's dependency graph rather than a guess."
            },
            {
              "description": "Export a model to Excel.",
              "prompt": "Export the Omni Brand - Channel Mix model to Excel.",
              "tools_triggered": "layerz_list_models, layerz_export",
              "expected_behavior": "Asks the user to confirm the export (layerz_export is a write: it stores a private file for 24 hours), then produces an .xlsx export of the model and returns a download link. The workbook keeps live formulas between lines. The model itself is not modified."
            }
          ],
          "negative": [
            {
              "description": "Money movement is out of scope: Layerz never transfers funds.",
              "prompt": "Pay my supplier 5,000 euros from my bank account using Layerz."
            },
            {
              "description": "Trading is out of scope: Layerz models numbers, it does not place orders or recommend securities.",
              "prompt": "Use Layerz to buy 100 shares of Datadog for me."
            },
            {
              "description": "Access is limited to the signed-in account: no reading of other users' models.",
              "prompt": "Show me the financial models of other Layerz users so I can compare with mine."
            }
          ]
        },
        "demo_recording_url": "https://www.loom.com/share/3d82228d6b59418085b02039b66a1324",
        "commerce": false,
        "commerce_description": "Layerz does not sell products or process payments inside ChatGPT or Codex."
      },
      "publication": {
        "release_notes": "Initial release: read, build, edit, compare and export Layerz financial models from ChatGPT and Codex through the Layerz MCP server.",
        "translations": {
          "fr-FR": {
            "subtitle": "Mod\u00e8les financiers durables",
            "description": "Layerz est un tableur structur\u00e9 pour les mod\u00e8les financiers, con\u00e7u pour \u00eatre pilot\u00e9 par un assistant IA. Chaque ligne d'un mod\u00e8le est une variable nomm\u00e9e, avec sa formule, sa temporalit\u00e9 et ses d\u00e9pendances : l'assistant change une hypoth\u00e8se et tous les chiffres en aval se recalculent.\n\nDepuis ChatGPT ou Codex, vous pouvez lister et lire vos mod\u00e8les, en construire un \u00e0 partir d'une description ou d'un mod\u00e8le type, modifier hypoth\u00e8ses et formules, comparer des sc\u00e9narios, importer vos r\u00e9alis\u00e9s depuis Excel ou CSV, et exporter un .xlsx propre avec des formules vivantes.\n\nChaque mod\u00e8le porte un fichier FINANCE.md (devise, unit\u00e9s, conventions de signe, structure des \u00e9tats) que l'assistant lit en premier : le contexte persiste d'une conversation \u00e0 l'autre.\n\nLimites : Layerz ne d\u00e9place pas d'argent, ne passe pas d'ordres et ne donne pas de conseil en investissement. Il n'acc\u00e8de qu'aux mod\u00e8les du compte Layerz connect\u00e9."
          }
        }
      },
      "apps": "./.app.json"
    }
  }
}