# Accretion Desk

> A merger model and accretion/dilution analysis that runs in the browser, followed by a paid review
> that challenges the assumptions and writes a merger consequences summary. Every number the review
> writes is checked against the model. https://accretion-desk.skillsafe.ai/

Accretion Desk is a web app on SkillSafe derived from the agent skill @anthropics/merger-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

- Offer price, premium, equity value, enterprise value, offer P/E and EV/EBITDA.
- Sources and uses: stock issued at the acquirer's current price for the stock share of the equity
  price; the rest of the uses (equity price, target debt refinanced, fees) paid from acquirer cash
  first, then new debt.
- Purchase price allocation: excess over book value, intangibles, deferred tax liability, goodwill
  and annual amortisation.
- Three-year pro forma EPS with synergy phase-in, on an adjusted basis (excluding new intangible
  amortisation, financing fee amortisation and one-time integration cost) and a GAAP basis.
- An EPS bridge from standalone to pro forma whose lines sum exactly to the change.
- Breakeven run-rate synergies for a Year 1 EPS-neutral deal, and the highest premium that keeps
  Year 1 neutral, on both bases.
- Sensitivity grids: premium by run-rate synergies for Years 1 to 3, and stock share by year.
- Flags: Year 1 or Year 3 dilution, accretion only with synergies, a P/E disadvantage on stock,
  premium outside 10% to 50%, leverage above 4x or 6x, new shares of 20% or more (the usual
  exchange shareholder-vote threshold), target holders at 50% or more, a bargain purchase, fees
  above 3% of enterprise value.

## What the review returns

One JSON object: `verdict` (supportable, stretched, not_supportable), `headline`, `consequences`,
`drivers` (bridge key, direction, reading), `challenges` (input field, severity, concern, test),
`flag_responses` (one per flag), `structure_options` (option, evidence from the grids, trade-off),
`diligence_questions`, `board_summary`, `summary`.

## Limits

It does not know market prices, consensus estimates or real companies. It does not model debt
paydown, buybacks, revenue synergies with their own margins, or purchase accounting beyond one
intangibles line. It is analysis, not investment advice.

## Links

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