# LBO Desk

> A leveraged buyout model that runs in the browser, followed by a paid review that challenges the
> assumptions and writes the returns case for an investment committee. Every number the review
> writes is checked against the model. https://lbo-desk.skillsafe.ai/

LBO Desk is a web app on SkillSafe derived from the agent skill @anthropics/lbo-model
(anthropics/financial-services-plugins, Apache-2.0). It runs on gpt-terra and is metered per review;
the model itself is free and needs no account.

## What the free model computes

- Sources and uses: purchase enterprise value (entry multiple x LTM EBITDA), transaction fees,
  financing fees and cash to the balance sheet, funded by a term loan, second lien and senior notes
  (each sized in turns of LTM EBITDA), rollover equity and sponsor equity as the plug.
- Operating model: revenue from a growth rate or a year-by-year path, an EBITDA margin moving in a
  straight line to the exit-year margin, D&A, capex, working capital as a share of revenue growth,
  and tax with losses carried forward.
- Debt schedule: interest on beginning balances (no circular reference), term loan amortisation,
  notes with cash or PIK interest, a revolver drawn when cash falls below the minimum, and a sweep
  of excess cash to the revolver, then the term loan, then the second lien. Balances never go
  negative.
- Exit: exit multiple x exit-year EBITDA less net debt and exit fees, a management pool, the
  sponsor's share, MOIC and IRR (single entry and exit cash flows).
- Returns bridge: EBITDA growth at the entry multiple, multiple change, net debt reduction and
  fees, summing exactly to the equity gain.
- Hurdle math: the highest entry multiple (and enterprise value) that clears the hurdle IRR, the
  exit multiple needed for the hurdle, and the exit multiple for 1.0x.
- Four 5x5 grids with the base case at the centre: IRR and MOIC by entry and exit multiple, IRR by
  revenue growth shift and exit multiple, IRR by total leverage and entry multiple.
- Model checks from the LBO verification checklist, and flags: IRR below the hurdle, MOIC under
  2.0x, reliance on multiple expansion, leverage above 5.0x or 6.5x, interest cover under 2.0x or
  1.5x, a liquidity shortfall, a revolver draw, equity under 30% of sources, margin expansion of 5
  points or more, revenue growth of 15% a year or more, slow deleveraging, PIK accrual, high fees.

## What the review returns

One JSON object: `verdict` (supportable, stretched, not_supportable), `headline`, `returns_case`,
`drivers` (bridge key, adds or subtracts, reading), `challenges` (input field, severity, concern,
test), `flag_responses` (one per flag), `structure_options` (option, evidence from the grids or
hurdle math, trade-off), `diligence_questions`, `ic_summary`, `summary`.

## Limits

It does not know market multiples, lender terms or real companies. It does not model dividend
recaps, add-ons, covenants, quarterly timing or purchase accounting. It is analysis, not
investment advice.

## Links

- App: https://lbo-desk.skillsafe.ai/
- API tutorial: https://lbo-desk.skillsafe.ai/api.html
- Source skill: https://skillsafe.ai/skill/@anthropics/lbo-model