Excelor Skill System Guide

Overview

Skills are markdown-based templates that define how to build financial models in Excel. They provide structured guidance for AI-generated spreadsheets, including layouts, formulas, formatting rules, and validation checks.

Skill Location

All skill files are stored in:

excelor/spreadsheets/skills/

Skill File Structure

Every skill follows this standard structure:

1. YAML Frontmatter (Required)

---
name: skill-name-with-dashes
description: Brief description of what this skill generates and when to use it.
---
FieldPurpose
nameMachine-readable identifier (lowercase, dashes)
descriptionHuman-readable explanation for skill selection

2. Markdown Body

The body contains detailed instructions organized into these sections:

SectionPurpose
OverviewWhat the model does and when to use it
Model ArchitectureSheet structure and layout specifications
Key FormulasCritical Excel formulas with explanations
Formatting StandardsColors, fonts, number formats
Validation ChecksChecklist to verify model correctness
Common TriggersKeywords that activate this skill

Skill Types

Single-Sheet Skills

Individual financial models on one sheet:

  • dcf_skill.md - DCF valuation
  • lbo_skill.md - LBO analysis
  • comps_model_skill.md - Comparable companies

Multi-Sheet Skills

Complex models with multiple interconnected sheets:

Overview Files (orchestrate the model):

  • ma_overview.md - M&A model index (10 sheets)
  • nav_overview.md - NAV model index (7 sheets)

Component Files (individual sheets):

  • ma_01_transaction_assumptions.md
  • ma_02_sources_uses.md
  • nav_01_assumptions.md
  • etc.

Industry-Specific Skills

Templates tailored for specific sectors:

  • three_statement_saas_skill.md
  • three_statement_bank_skill.md
  • three_statement_retail_skill.md
  • healthcare_pharma_skill.md

How Skills Are Used

  1. Skill Selection: The AI matches user requests to skills via "Common Triggers"
  2. Skill Loading: The get_skill tool fetches the skill content
  3. Model Generation: AI follows the skill's architecture and formulas
  4. Validation: AI runs through the validation checklist

Example Workflow

User: "Build an LBO model for Target Corp"

1. AI matches "LBO model" to lbo_skill.md
2. AI calls: get_skill("lbo-model-generation")
3. AI follows skill's Model Architecture section
4. AI creates sheets: Assumptions, Operating Model, Debt Schedule, Returns
5. AI applies formatting from Formatting Standards
6. AI verifies using Validation Checks

Current Skills Inventory

CategoryCountExamples
Valuation8DCF, DDM, Comps, APV, Real Options
LBO/PE2LBO, Take-Private
M&A11Transaction Assumptions, Sources & Uses, PPA, Synergies, Accretion/Dilution
NAV/REIT8Assumptions, Asset Schedule, Liabilities, Premium/Discount
Three-Statement7SaaS, Bank, Retail, Healthcare, Insurance, Energy, Telecom
Credit/Risk5Credit Analysis, Debt Capacity, Merton Model, VaR, Altman Z-Score
Specialty7SaaS Metrics, Startup Valuation, IPO Valuation, Bank FIG

Total: 48 skill files

Related Documentation