Airsc miễn phí cho tất cả. Nếu nó giúp bạn, hãy ủng hộ $2–$20.

Ủng hộ
PluginCẩn thậnDữ liệu & phân tíchCần cài đặt chút

excel-analyst-pro-skill-md

jeremylongshore · jeremylongshore/excel-analyst-pro-skill-md

Kỹ năng này cho phép Claude Code tạo các mô hình tài chính chuyên nghiệp trong Excel, bao gồm định giá DCF, phân tích LBO và báo cáo phương sai.

Professional financial modeling toolkit for Claude Code with auto-invoked Skills. Build DCF models, LBO analysis, variance reports, and pivot tables using natural language.

57 sao5Giấy phép: Không có giấy phépCập nhật tháng trước0 lượt mở

Xem trước

▍

Bạn có thể làm gì với nó

  • Xây dựng mô hình Dòng tiền chiết khấu (DCF) năm năm từ các dự báo tài chính và lịch trình nợ đã cung cấp.
  • Thực hiện phân tích lợi nhuận Mua lại bằng nợ (LBO) cho các kịch bản đầu tư.
  • Tạo sổ làm việc phân tích phương sai ngân sách, dự báo hoặc kỳ.

Kiểm tra an toàn

Cẩn thận

Dùng được, nhưng nên đọc ghi chú trước.

  • ●Không có giấy phép rõ ràng: hỏi tác giả trước khi dùng cho sản phẩm thương mại.

Quét tự động các file trong repo. Luôn xem lại code trước khi chạy.

Trích từ README

Excel Analyst Pro

Excel Analyst Pro turns supplied financial evidence into reviewable Excel workbooks for three jobs:

  • DCF valuation
  • LBO returns analysis
  • Budget, forecast, or period variance analysis

It shows its assumptions, links formulas to inputs, checks the saved workbook, and keeps spreadsheet-engine recalculation status explicit. Its read-only verifier checks whether supplied evidence is internally consistent; it does not pretend that a self-hash authenticates Excel or LibreOffice. The workflow does not invent company data, root causes, or financing terms.

Understand it in 30 seconds

Ask for the workbook and provide the evidence you have:

Build a five-year DCF from this forecast and debt schedule. Save a new workbook;
do not overwrite the source. Mark every assumption I still need to approve.

The skill will:

  1. confirm the decision, source files, units, periods, and output path;
  2. separate supplied facts, sourced facts, assumptions, formulas, and gaps;
  3. build only the requested DCF, LBO, or variance workbook;
  4. verify formulas, balances, preservation limits, and recalculation status; and
  5. return the file path with a short evidence and verification receipt.

Native pivot tables, charts, slicers, What-If Data Tables, and other unsupported Excel objects are not promised. The skill can offer a formula-backed summary or an implementation plan when the active spreadsheet tool lacks a required feature.

Install

Clone the public repository, then load it directly. This does not require a marketplace registration:

git clone https://github.com/jeremylongshore/excel-analyst-pro-skill-md.git
claude --plugin-dir ./excel-analyst-pro-skill-md

The plugin includes one model-neutral AgentSkills.io entrypoint at skills/excel-analyst-pro/SKILL.md. Codex and other AgentSkills-compatible harnesses can install or link that skill directory through their normal skill installer; the workflow itself does not call a model API.

The Claude plugin does not start or download an MCP server automatically. An opt-in example for @negokaz/excel-mcp-server@0.12.0 is provided at examples/claude-mcp.json. It is a complete mcpServers configuration object; review it and explicitly merge it into your own Claude configuration only if you consent to npx --yes downloading and executing that pinned package. For stronger supply-chain control, install the audited package into a locked local tool directory and replace npx --yes with its fixed local executable path.

The adapter's published README requires Node.js 20 or later with npm/npx. Its package.json does not enforce that requirement with an engines field, so deployment must check it explicitly. The adapter can read and write ordinary sheets, values, formulas, tables, and direct cell styles. It is not a native Excel automation surface and does not recalculate formulas.

The bundled launch command is the adapter's macOS/Linux npx form. Its published Windows configuration launches npx through cmd /c; Windows operators must apply that platform-specific command before enabling the adapter.

The packaged inventory, CSV reconciliation, and artifact-verification helpers use only the Python standard library and require Python 3.10 or later. The model workflow remains usable without either optional runtime, but it must stop instead of claiming an artifact operation it cannot perform.

Safety and evidence

  • Existing workbooks are edited through a copy unless in-place work is explicit.
  • Overwrites require confirmation.
  • Unsupported macros, links, names, validations, protection, or other workbook features block mutation when preservation matters.
  • Formula output is not called verified until a compatible engine recalculates the saved workbook and key cells are reopened and checked.
  • Missing evidence stays missing; static industry defaults are not substituted.
  • Variance drivers remain hypotheses unless source data proves them.

Migration from v1.1

Version 2 is a deliberate consolidation:

Previous entrypointVersion 2 destination
excel-dcf-modelerexcel-analyst-pro, DCF mode
excel-lbo-modelerexcel-analyst-pro, LBO mode
excel-variance-analyzerexcel-analyst-pro, variance mode
excel-pivot-wizardSuspended; no unsupported native pivot/chart claim
/build-dcf, /build-lbo, /analyze-varianceUse natural language or $excel-analyst-pro

The former files remain recoverable in Git history. Existing prompts continue to work when they clearly request a DCF, LBO, or variance workbook, but automation should update explicit skill names to excel-analyst-pro.

Package layout

.claude-plugin/plugin.json
examples/claude-mcp.json
skills/excel-analyst-pro/
├── SKILL.md
├── agents/openai.yaml
├── scripts/
│   ├── profile_reconcile.py
│   ├── verify_artifact.py
│   └── workbook_inventory.py
└── references/
    ├── artifact-contract.md
    ├── dcf.md
    ├── lbo.md
    ├── tooling.md
    └── variance.md

Intent Solutions Proprietary. See LICENSE.

Xem mã nguồn trên GitHub