← ClaudeAtlas

kelly-financelisted

Build, analyze, repair, and audit finance three-statement models and operating forecasts. Use when the user invokes $kelly-finance, asks for 财务三表, 三表模型, income statement, balance sheet, cash flow statement, forecast model, budget model, financial projections, FP&A model, SaaS/unit economics forecast, fundraising model, board finance pack, investor model, model checks, balance-sheet balancing, cash runway, working capital, capex/depreciation schedules, debt schedules, scenario/sensitivity cases, or Excel/Sheets-ready finance outputs.
mr-kelly/skills · ★ 3 · AI & Automation · score 74
Install: claude install-skill mr-kelly/skills
# Kelly Finance ## Overview Use this skill as a practical FP&A and corporate-finance modeling desk. It builds clean three-statement models, turns assumptions into forecast tables, audits model logic, and explains finance outputs in English or Chinese. For spreadsheet file creation or editing, also use the local spreadsheet/xlsx tooling if available. This skill owns the finance logic, modeling structure, checks, and presentation standard. Default interaction mode: App UI. Unless the user explicitly asks for chat-only handling, generate or load the local model snapshot, start/reuse the local app with `app/start.sh`, and give the actual local URL. Use chat-only mode only when the user says "纯聊天", "chat only", "不要打开 UI", or similar. ## Boundary - The skill may build local workbooks, inspect local workbook/export files, normalize assumptions, write local model snapshots, and draft model-review recommendations. - The app reads and writes local files only. It must not connect to banks/accounting systems, send files, mutate ERP records, move money, or change external systems. - Treat financial models and assumptions as sensitive. Do not commit `config.local.json`, env files, `app/.data/`, private workbook exports, statements, customer/vendor data, or execution reports. - Any external action, such as sending a model to investors or changing source-of-truth books, is approval-required and executed outside the app by the agent after human review. ## Local App Start the model rev