@v587d/capital-generation 2.7.7 → 2.7.8

package/selected-skills/financeskills/skills/corporate-consolidation/SKILL.md

@@ -0,0 +1,84 @@
+---
+name: corporate-consolidation
+description: When the user wants to consolidate the financial results of multiple legal entities into a single set of statements. Also use when the user mentions "intercompany eliminations," "minority interest," "NCI calculation," "equity method," "business combinations," or "IFRS 10 / ASC 810."
+metadata:
+  version: 1.0.0
+---
+
+# Corporate Consolidation
+
+You are a Group Controller. Your goal is to combine the financial results of a parent company and its subsidiaries as if they were a single economic entity.
+
+## Initial Assessment
+
+1. **Ownership Structure**
+   - Which entities are controlled (>50% ownership)?
+   - Which entities use the Equity Method (20-50% ownership)?
+   - Are there Non-Controlling Interests (NCI)?
+
+2. **Intercompany Transactions**
+   - Are there intercompany sales, loans, or management fees?
+   - Is there "unrealized profit" in inventory (e.g., Parent sold to Sub but Sub hasn't sold to an external party)?
+
+3. **Currency Conversion**
+   - Do subsidiaries use a different functional currency?
+
+---
+
+## Consolidation Framework
+
+### Priority Order
+1. **Standardization** (Ensuring all subsidiaries use the same accounting policies).
+2. **Translation** (Converting sub-results to the Parent's reporting currency).
+3. **Aggregation** (Adding line items together).
+4. **Eliminations** (Removing intercompany balances and equity investments).
+5. **NCI Calculation** (Allocating profit and equity to minority owners).
+
+---
+
+## Technical Consolidation Steps
+
+### 1. Intercompany Eliminations
+- **Balance Sheet**: Eliminate intercompany Receivables vs. Payables.
+- **Income Statement**: Eliminate intercompany Sales vs. COGS.
+
+### 2. The Investment Elimination
+- Debit: Subsidiary Equity accounts (Common Stock, RE).
+- Credit: Parent's "Investment in Subsidiary" asset account.
+- Plug: Goodwill or Bargain Purchase Gain.
+
+### 3. Non-Controlling Interest (NCI)
+- Allocate a percentage of the subsidiary's net income and equity to external owners.
+
+---
+
+## Output Format
+
+### Consolidated Financial Package
+
+**Consolidation Worksheet**
+- Columns: Parent | Sub A | Sub B | Eliminations | Consolidated Total.
+
+**Consolidated Statements**
+- P&L, Balance Sheet, and Cash Flow reflecting the whole group.
+
+**Intercompany Log**
+- Detailed list of eliminated transactions for audit verification.
+
+---
+
+## Scripts
+- [calculate.py](./scripts/calculate.py): Deterministic functions for this skill's core computations. Run `python3 scripts/calculate.py` to self-test; import the functions instead of doing mental math.
+
+---
+
+## References
+- [IFRS 10 Summary](./references/consolidation-standard.md): Control and consolidation rules.
+- [NCI Accounting](./references/minority-interest.md): How to calculate and present NCI.
+
+---
+
+## Related Skills
+- **financial-statement-prep**: Consolidation is the final step in group reporting.
+- **audit-checklist**: For auditing the complex elimination entries.
+- **ecl-computation**: For assessing credit risk on a consolidated basis.

package/selected-skills/financeskills/skills/corporate-consolidation/references/consolidation-standard.md

@@ -0,0 +1,28 @@
+# Consolidation Standards: IFRS 10 / ASC 810 Essentials
+
+## Control Test (IFRS 10)
+Control = power over relevant activities + exposure to variable returns + ability to use power to affect returns. Consolidate ALL controlled entities; >50% voting rights is the usual but not the only path (de facto control, potential voting rights, structured entities).
+
+## Method Selection
+| Holding/Influence | Method |
+|---|---|
+| Control (usually >50%) | Full consolidation |
+| Joint control — joint venture | Equity method (IFRS 11) |
+| Significant influence (20%–50%) | Equity method |
+| < 20%, no influence | FVTPL/FVOCI financial asset |
+
+## Consolidation Mechanics
+1. Combine 100% of subsidiary line items (even at < 100% ownership).
+2. Eliminate parent's investment against subsidiary equity at acquisition; difference = goodwill (or bargain purchase gain).
+3. Recognize NCI (see minority-interest.md).
+4. Eliminate ALL intragroup balances, transactions, and unrealized profits (inventory, fixed assets).
+5. Align accounting policies and reporting dates (gap ≤ 3 months with adjustments).
+6. Translate foreign subs: assets/liabilities at closing rate, income at average rate, equity at historical; differences → FX translation reserve (OCI/CTA).
+
+## Business Combination (IFRS 3) at Acquisition
+- Identifiable assets/liabilities at fair value (including intangibles not on the target's books: customer relationships, brands).
+- Contingent consideration at fair value; subsequent changes through P&L (if liability-classified).
+- Transaction costs expensed, NOT capitalized.
+
+## Loss of Control
+Derecognize sub, recognize retained stake at fair value, recycle CTA, gain/loss to P&L.

package/selected-skills/financeskills/skills/corporate-consolidation/references/minority-interest.md

@@ -0,0 +1,25 @@
+# Non-Controlling Interest (NCI)
+
+## Measurement Choice at Acquisition (per transaction, IFRS 3)
+- **Fair value method ("full goodwill")**: NCI at fair value → goodwill includes NCI's share. Required under US GAAP.
+- **Proportionate method ("partial goodwill")**: NCI = NCI% × identifiable net assets at fair value → goodwill is parent's share only. IFRS option.
+
+Worked example — Parent buys 80% for 800; identifiable net assets FV = 700; NCI fair value = 190:
+- Full goodwill: (800 + 190) − 700 = **290** (NCI = 190)
+- Partial goodwill: 800 − 80% × 700 = **240** (NCI = 20% × 700 = 140)
+
+## Subsequent Measurement
+`NCI closing = NCI opening + NCI% × sub's comprehensive income − dividends paid to NCI ± NCI share of equity transactions`
+
+Allocate losses to NCI even if it goes negative (IFRS 10).
+
+## Profit Attribution
+P&L shows single consolidated profit, then attribution: "attributable to owners of the parent" and "attributable to NCI". NCI share is of the subsidiary's PAT adjusted for fair-value depreciation, unrealized profit eliminations where the sub is the seller (upstream), and intragroup interest.
+
+## Changes in Ownership Without Loss of Control
+Equity transactions — no P&L gain/loss, no goodwill remeasurement. Difference between consideration and NCI adjustment goes to parent equity.
+
+## Common Errors
+1. Eliminating only the parent's share of intragroup profit (must eliminate 100%).
+2. Forgetting fair-value adjustment depreciation in NCI's profit share.
+3. Recording gain on buying out NCI (it's an equity transaction).

package/selected-skills/financeskills/skills/credit-analysis/EXAMPLE.md

@@ -0,0 +1,21 @@
+# Worked Example - credit-analysis
+
+## Ask
+
+> Borrower requests a 500M term loan, 5 years. EBITDA 280M (includes a 40M one-off insurance recovery), cash taxes 45M, maintenance capex 60M, existing debt 320M with 38M annual interest, new loan interest ~12%, straight-line amortization. Assess debt service capacity.
+
+## What a correct response contains
+
+Normalize EBITDA to 240M. Total debt 820M -> leverage 3.4x. Total interest ~98M, principal 100M/yr. DSCR = (240-45-60)/(98+100) = 135/198 = ~0.68x - fails. Restructure needed: smaller amount, longer tenor, or back-ended amortization.
+
+## File-driven variant
+
+With fixture(s) `evals/files/borrower_financials.csv`:
+
+> Assess the borrower in the attached file (evals/files/borrower_financials.csv): compute interest coverage, net leverage, and DSCR, then measure headroom against both covenants and state which is tighter.
+
+Expected: Interest coverage 4.0x; net leverage 3.2x vs max 4.0x (20% headroom); DSCR 1.20x vs min 1.25x (-4.0% headroom). DSCR is the binding constraint.
+
+## Compute, don't estimate
+
+All figures above come from [`scripts/calculate.py`](./scripts/calculate.py) - import its functions rather than doing arithmetic in prose.

package/selected-skills/financeskills/skills/credit-analysis/SKILL.md

@@ -0,0 +1,84 @@
+---
+name: credit-analysis
+description: When the user wants to assess a borrower's creditworthiness, debt capacity, or covenant compliance. Also use when the user mentions "credit memo," "debt service coverage," "leverage ratio," "covenant headroom," "5 Cs of credit," "credit rating," "loan underwriting," or "can they repay."
+metadata:
+  version: 1.0.0
+---
+
+# Credit Analysis
+
+You are a Credit Analyst. Your goal is to answer one question with evidence: will this borrower service and repay its obligations through the cycle — and what protects the lender if not?
+
+## Initial Assessment
+
+1. **Exposure Context**
+   - Facility type (term, revolver, project, trade), amount, tenor, seniority, security.
+   - Purpose of funds (the second C — speculative purposes change everything).
+
+2. **Information Set**
+   - ≥ 3 years financials (audited preferred; note qualifications), projections, bank statements, aging reports, existing debt schedule with covenants.
+
+---
+
+## Credit Framework
+
+### Quantitative Core
+| Ratio | Formula | Typical Comfort (corporate) |
+|---|---|---|
+| Leverage | Total debt / EBITDA | < 3.0x (cyclicals lower) |
+| Interest coverage | EBITDA / Interest | > 3.0x |
+| DSCR | (EBITDA − cash taxes − maintenance capex) / (interest + scheduled principal) | > 1.2–1.3x |
+| FCF/debt | FCF / Total debt | > 10% |
+| Liquidity | (Cash + undrawn committed) / 12-mo obligations | > 1.0x |
+Compute on a NORMALIZED basis (strip one-offs) and on a stressed basis.
+
+### Debt Capacity
+`Max debt ≈ min(target leverage × normalized EBITDA, debt serviceable at stressed DSCR floor)` — the binding constraint is usually the stressed DSCR, not the leverage multiple.
+
+### Qualitative (the part models miss)
+Management quality and track record through downturns; industry position and cyclicality; customer/supplier concentration (> 20% single name = structural risk); FX mismatch between revenue and debt currency (the classic emerging-market killer); group structure and structural subordination.
+
+### Security & Recovery
+Collateral valuation with haircuts (real estate 20–30%, inventory 50%+, receivables by aging quality), perfection status verified (registrations!), guarantee strength (guarantor's own credit), and estimated recovery in liquidation vs. going concern.
+
+---
+
+## Technical Analysis Steps
+
+1. **Compute every ratio from the formula stated above** across history + projection, writing out the arithmetic; trend matters more than the level.
+2. **Stress test**: revenue −15–25%, margin −200–400 bps, rates +300 bps, FX devaluation scenario for currency-mismatched borrowers; DSCR must stay ≥ 1.0 in the moderate stress or the structure needs work (more equity, shorter tenor, cash sweep, reserves).
+3. **Covenant design/compliance**: headroom analysis (covenant level vs. projection ≥ 20–25% cushion), definitions audit (what's in "EBITDA"? addback abuse), compliance certificate recalculation — never accept the borrower's spreadsheet.
+
+---
+
+## Output Format
+
+### Credit Memorandum
+
+**Recommendation**: approve/decline/structure, amount, tenor, pricing guidance, conditions precedent.
+
+**Borrower Analysis**: business, management, industry, the 5 Cs evidence.
+
+**Financial Analysis**: ratio table (historic/projected/stressed) with every figure traced to the formula and inputs that produced it.
+
+**Risks & Mitigants**: ranked, each with a real mitigant (a "strong sponsor" is not a mitigant without a keepwell).
+
+**Security & Documentation**: collateral, covenants proposed with headroom rationale.
+
+---
+
+## Scripts
+- [calculate.py](./scripts/calculate.py): Deterministic functions for this skill's core computations, kept as reviewable material. This plugin's research role has no shell and the scripts are not part of the installed package, so nothing here runs them: work the ratio definitions as written above and in [Covenant Design Guide](./references/covenant-guide.md), show each calculation in the answer, and hand anything that needs real compute to the delegated `data_junior`.
+
+---
+
+## References
+- [Covenant Design Guide](./references/covenant-guide.md): Covenant selection, definitions hygiene, and headroom setting.
+
+---
+
+## Related Skills
+- **financial-analysis**: The ratio engine this skill builds on.
+- **ecl-computation**: Portfolio-level expected loss on the exposure.
+- **debt-restructuring**: When the answer to "can they repay" is no.
+- **risk-assessment**: Enterprise risk context of the borrower.

package/selected-skills/financeskills/skills/credit-analysis/references/covenant-guide.md

@@ -0,0 +1,26 @@
+# Covenant Design Guide
+
+## Covenant Taxonomy
+- **Maintenance** (tested quarterly): leverage, coverage, DSCR, minimum net worth, minimum liquidity. Bank-market standard.
+- **Incurrence** (tested on action): limits on additional debt, dividends/restricted payments, asset sales, liens, investments. Bond-market standard; covenant-lite loans import this style.
+- **Information**: monthly/quarterly reporting, compliance certificates, budget delivery, auditor access.
+
+## The Big Four Financial Covenants
+| Covenant | Definition Discipline |
+|---|---|
+| Net leverage | Cap EBITDA addbacks (run-rate synergies ≤ 10–20% of EBITDA, time-boxed); define netted cash (exclude restricted) |
+| Interest coverage | Cash interest, not effective-rate accounting interest; include capitalized interest |
+| DSCR | Include scheduled amortization AND mandatory prepayments; exclude voluntary |
+| Minimum liquidity | Committed undrawn counts only while conditions to draw are met |
+
+## Headroom Setting
+Set covenants off the BORROWER's base case with 20–25% EBITDA cushion at the tightest point, widening over time only with deleveraging. Headroom < 15% = expect waiver requests; > 40% = covenant is decoration. Model the cushion quarterly, not annually — seasonality breaches happen mid-year.
+
+## Definitions Hygiene (where deals are won/lost)
+1. "EBITDA" addbacks: exceptional items capped and itemized; pro-forma acquisitions limited to completed deals with diligence-quality numbers.
+2. "Debt": include guarantees, leases (state convention), deferred consideration, factoring with recourse.
+3. Equity cure: limit frequency (2–3 over life, not consecutive), cash must actually arrive, no over-cure gaming.
+4. FX: covenant currency and testing-date rates specified; high-volatility currencies need average-rate testing or you covenant FX noise.
+
+## Early Warning Stack (monitor between test dates)
+Utilization creep on the revolver, aged payables stretching, management account delays, auditor change, key customer loss, covenant-adjacent metrics (e.g., 12.1x against a 12.5x cap two quarters running). A covenant breach is a LAGGING indicator — the file should show concern before the certificate does.

package/selected-skills/financeskills/skills/crypto-accounting/EXAMPLE.md

@@ -0,0 +1,21 @@
+# Worked Example - crypto-accounting
+
+## Ask
+
+> We bought 1 BTC for $50k. It dropped to $30k last month, but now it's at $60k. How do we value it on our GAAP balance sheet?
+
+## What a correct response contains
+
+Should identify the GAAP impairment rule (Indefinite-lived intangible asset). Should explain that the asset must be written down to $30k and cannot be written back up to $60k. Should provide a Digital Asset Ledger Summary.
+
+## File-driven variant
+
+With fixture(s) `evals/files/btc_lots.csv`:
+
+> Holdings per the attached lot file (evals/files/btc_lots.csv), reporting under US GAAP (ASU 2023-08). (a) Year-end price is $36,000/unit and prior carrying was $110,000 total: compute the fair-value remeasurement P&L. (b) Separately, if 2.5 units are sold at $70,000 with FIFO cost relief, compute the realized gain.
+
+Expected: (a) New carrying $108,000, P&L -2,000. (b) FIFO cost basis $85,000, realized gain $90,000.
+
+## Compute, don't estimate
+
+All figures above come from [`scripts/calculate.py`](./scripts/calculate.py) - import its functions rather than doing arithmetic in prose.

package/selected-skills/financeskills/skills/crypto-accounting/SKILL.md

@@ -0,0 +1,81 @@
+---
+name: crypto-accounting
+description: When the user wants to account for digital assets (Cryptocurrency, NFTs) in financial statements. Also use when the user mentions "stablecoin reconciliation," "crypto impairment," "digital asset classification," "accounting for Bitcoin," or "gas fee accounting."
+metadata:
+  version: 1.0.0
+---
+
+# Crypto-Accounting & Digital Assets
+
+You are a Digital Asset Controller. Your goal is to apply existing accounting frameworks (like IAS 38 or ASC 350) to the unique properties of blockchain-based assets.
+
+## Initial Assessment
+
+1. **Asset Classification**
+   - **Intangible Asset**: Standard for held-for-investment (e.g., BTC, ETH).
+   - **Inventory**: If the company is a broker-dealer in digital assets.
+   - **Financial Asset**: Rare (usually reserved for assets with contractual rights to cash).
+
+2. **Measurement Model**
+   - **Cost Model**: Value at cost, test for impairment (most common under GAAP).
+   - **Revaluation Model**: Fair value (allowed under IFRS for active markets).
+
+3. **Transaction Context**
+   - Staking rewards, Mining income, Gas fees, Airdrops.
+
+---
+
+## Crypto Framework
+
+### Priority Order
+1. **Identification & Ownership** (Verifying control via private keys/custody).
+2. **Cost Basis Tracking** (FIFO or Specific ID for every transaction).
+3. **Impairment Testing** (Checking if the market price dropped below book value).
+4. **Income Recognition** (Determining the value of rewards at the moment of receipt).
+5. **Reconciliation** (Matching on-chain activity to the General Ledger).
+
+---
+
+## Technical Accounting Steps
+
+### 1. Impairment (GAAP Standard)
+- If the price of Bitcoin drops at any point during the quarter, write down the asset to the lowest point. 
+- You cannot write it back up under current US GAAP (unless it's sold and repurchased).
+
+### 2. Gas Fee Accounting
+- Generally treated as a transaction cost. Capitalize if part of an acquisition; expense if for a transfer.
+
+---
+
+## Output Format
+
+### Digital Asset Ledger Summary
+
+**Inventory Table**
+- Asset | Quantity | Cost Basis | Current FMV | Carrying Value.
+
+**Activity Log**
+- Staking Income recognized.
+- Realized Gains/Losses from disposals.
+- Impairment charges taken.
+
+**Wallet Reconciliation**
+- Discrepancy report between wallet balance and GL balance.
+
+---
+
+## Scripts
+- [calculate.py](./scripts/calculate.py): Deterministic functions for this skill's core computations. Run `python3 scripts/calculate.py` to self-test; import the functions instead of doing mental math.
+
+---
+
+## References
+- [IFRS Digital Asset Guide](./references/ifrs-crypto.md): Intangible vs Inventory.
+- [FASB Crypto Update](./references/fasb-crypto.md): New rules on fair value reporting.
+
+---
+
+## Related Skills
+- **ai-anomaly-detection**: To detect unusual wallet activity.
+- **automated-reconciliation**: For matching on-chain data to the ledger.
+- **financial-statement-prep**: For accurately reporting the "Digital Assets" line item.

package/selected-skills/financeskills/skills/crypto-accounting/references/fasb-crypto.md

@@ -0,0 +1,26 @@
+# FASB Crypto Accounting: ASU 2023-08 (ASC 350-60)
+
+## Scope Test (all six required)
+1. Meets the definition of an intangible asset;
+2. Holder has no enforceable claims on underlying goods/services/other assets (excludes most stablecoins and wrapped/asset-backed tokens);
+3. Created/resides on a distributed ledger;
+4. Secured by cryptography;
+5. Fungible (excludes NFTs);
+6. Not created by the reporting entity or its related parties (excludes self-issued tokens).
+
+## Measurement
+**Fair value through net income, every reporting period** (ASC 820). This replaced the old cost-less-impairment model — both gains AND losses now flow through P&L. Effective fiscal years beginning after 15 Dec 2024; early adoption was permitted.
+
+## Presentation & Disclosure
+- Balance sheet: crypto assets presented SEPARATELY from other intangibles.
+- Income statement: remeasurement gains/losses presented separately from impairments of other intangibles.
+- Disclosures: per significant holding — name, cost basis, fair value, units; rollforward (additions, dispositions, gains, losses); restrictions; method for cost basis (FIFO/specific ID).
+
+## Out-of-Scope Items (still messy)
+- **Stablecoins**: fail criterion 2 → financial asset analysis; often amortized cost if redeemable 1:1.
+- **NFTs**: remain under legacy intangible model (cost less impairment) — non-fungible.
+- **Self-issued tokens**: no recognized asset for the issuer; proceeds analysis → liability/revenue/equity question under contract terms.
+- **Staking rewards**: recognize when control obtained, at fair value — revenue vs. other income classification judgment.
+
+## IFRS Contrast
+No crypto-specific standard: IAS 38 (cost or revaluation-through-OCI if active market) or IAS 2 at FV-less-costs-to-sell for broker-traders. The US GAAP FV-through-P&L answer is currently the cleaner model — see ifrs-crypto.md.

package/selected-skills/financeskills/skills/crypto-accounting/references/ifrs-crypto.md

@@ -0,0 +1,25 @@
+# IFRS Accounting for Digital Assets
+
+Guidance on how to apply International Financial Reporting Standards to cryptocurrencies and other digital assets.
+
+## 🏛️ Classification Logic
+
+### 1. Intangible Assets (IAS 38)
+The standard treatment for cryptocurrencies like Bitcoin.
+- **Criteria**: Identifiable non-monetary asset without physical substance.
+- **Measurement**: Cost model or Revaluation model (if an active market exists).
+
+### 2. Inventory (IAS 2)
+Applicable if the entity holds the assets for sale in the ordinary course of business (e.g., broker-dealers).
+- **Measurement**: Lower of cost and net realizable value (NRV).
+
+---
+
+## 🔍 Key Challenges
+- **Impairment**: Determining if a price drop is "prolonged" or "material."
+- **Custody**: Verifying that the entity has control over the private keys.
+- **Valuation**: Selecting a representative exchange price (Level 1 input).
+
+## 🔗 Sources
+- [IFRS Interpretations Committee: Holdings of Cryptocurrencies](https://www.ifrs.org/projects/completed-projects/2019/holdings-of-cryptocurrencies/)
+- [PwC: Crypto-assets: IFRS reporting](https://www.pwc.com/gx/en/audit-assurance/publications/ifrs-reporting-crypto-assets.html)

package/selected-skills/financeskills/skills/cvp-breakeven/EXAMPLE.md

@@ -0,0 +1,21 @@
+# Worked Example - cvp-breakeven
+
+## Ask
+
+> Price 1,200, variable cost 780, fixed costs 8,400,000/year. How many units to break even, and how many for 2,100,000 after-tax profit at 30% tax?
+
+## What a correct response contains
+
+Contribution 420/unit. Breakeven 20,000 units (24,000,000 revenue). After-tax target needs pre-tax 3,000,000 -> (8,400,000+3,000,000)/420 = 27,143 units.
+
+## File-driven variant
+
+With fixture(s) `evals/files/product_economics.csv`:
+
+> From the attached unit economics (evals/files/product_economics.csv): compute breakeven units and sales, units needed for the target profit, margin of safety at current volume, and degree of operating leverage.
+
+Expected: Breakeven 6,000 units ($300,000); target requires 8,000 units; MOS 25%; DOL 4.0.
+
+## Compute, don't estimate
+
+All figures above come from [`scripts/calculate.py`](./scripts/calculate.py) - import its functions rather than doing arithmetic in prose.

package/selected-skills/financeskills/skills/cvp-breakeven/SKILL.md

@@ -0,0 +1,78 @@
+---
+name: cvp-breakeven
+description: When the user wants cost-volume-profit analysis, breakeven points, or operating leverage assessment. Also use when the user mentions "breakeven," "contribution margin," "margin of safety," "operating leverage," "how many units to be profitable," "target profit volume," or "should we accept this order below full cost."
+metadata:
+  version: 1.0.0
+---
+
+# Cost-Volume-Profit & Breakeven Analysis
+
+You are a Management Accountant. Your goal is to expose how profit responds to volume, price, and cost structure — and to use that to answer real decisions: pricing floors, special orders, capacity steps, and risk.
+
+## Initial Assessment
+
+1. **Cost Structure Split**
+   - Separate fixed from variable honestly (regression/high-low on history beats labels — "salaries" often contains variable overtime; "logistics" contains fixed contracts).
+   - Identify STEP costs (supervision, warehouse space, license tiers) and their breakpoints.
+
+2. **Decision Context**
+   - Single product, multi-product mix, special order, or capacity decision — each uses different math.
+
+---
+
+## CVP Framework
+
+### Core Formulas
+- `Contribution per unit = Price − Variable cost per unit`
+- `CM ratio = Contribution / Price`
+- `Breakeven units = Fixed costs / contribution per unit`
+- `Breakeven revenue = Fixed costs / CM ratio`
+- `Target profit volume = (Fixed + target profit) / contribution per unit` (pre-tax; for after-tax: target/(1−t))
+- `Margin of safety % = (Actual − breakeven revenue)/Actual`
+- `Operating leverage = Contribution / Operating income` — at DOL 4, a 10% volume drop cuts profit 40%.
+
+### Multi-Product Mix
+Weighted-average contribution at the CONSTANT MIX assumption: `BE units = Fixed / Σ(mix% × unit contribution)`. State the assumption loudly — mix shift is the usual reason "we hit breakeven volume but lost money."
+
+### Special Order / Pricing Floor Logic
+Accept below full cost IF: price > variable cost + order-specific incremental costs, spare capacity exists (else add opportunity cost of displaced contribution), and no market-spoilage effect (price leakage to existing customers). Full-cost allocations are irrelevant to the decision (product-profitability CM hierarchy).
+
+---
+
+## Technical Analysis Steps
+
+1. **Validate the cost split with code**: regress total cost on volume (slope = variable rate, intercept = fixed proxy); R² < 0.7 → segment the data or find the second driver.
+2. **Breakeven with steps**: piecewise fixed costs create MULTIPLE breakeven points — chart profit vs. volume across the full range including the steps.
+3. **Risk view**: margin of safety, DOL, and a sensitivity table (price ±5%, volume ±10%, variable cost ±5%) — price moves dominate; show why (price flows straight to contribution).
+
+---
+
+## Output Format
+
+### CVP Analysis
+
+**Cost Structure**: fixed/variable/step split with derivation basis.
+
+**Breakeven**: units and revenue, with the profit-volume chart description (including steps).
+
+**Decision Answer**: the specific question answered with incremental logic shown.
+
+**Risk Profile**: margin of safety, operating leverage, sensitivity table.
+
+---
+
+## Scripts
+- [calculate.py](./scripts/calculate.py): Deterministic functions for this skill's core computations. Run `python3 scripts/calculate.py` to self-test; import the functions instead of doing mental math.
+
+---
+
+## References
+- [CVP Decision Patterns](./references/cvp-patterns.md): Worked special-order, step-cost, and mix-shift examples.
+
+---
+
+## Related Skills
+- **product-profitability**: The contribution hierarchy feeding CVP.
+- **variance-analysis**: Explaining actual vs. CVP-predicted profit.
+- **budget-forecast**: CVP as the engine of flexible budgets.
+- **pricing decisions**: Pair with unit economics for pricing floors.

package/selected-skills/financeskills/skills/cvp-breakeven/references/cvp-patterns.md

@@ -0,0 +1,28 @@
+# CVP Decision Patterns (worked)
+
+## 1. Basic Breakeven With Target Profit
+Price 500, variable 320, fixed 5,400,000. Contribution 180; CM ratio 36%.
+- Breakeven = 30,000 units / 15,000,000 revenue.
+- After-tax target 1,260,000 at 30% tax → pre-tax 1,800,000 → volume = (5,400,000+1,800,000)/180 = **40,000 units**.
+
+## 2. Step Cost Trap
+Same product, but each 10,000 units of capacity needs a supervisor (600,000).
+Profit at 30,000 units with 3 supervisors: contribution 5,400,000 − fixed 5,400,000 − 1,800,000 = **−1,800,000 (loss!)** — the naive breakeven ignored the steps. Recompute piecewise: breakeven actually lands at 40,000 units (4 supervisors: (5,400,000+2,400,000)/180 = 43,333 → check against the 40,001–50,000 band: feasible at 43,333). Always test the breakeven against the step band it falls in.
+
+## 3. Special Order
+Capacity 100,000; running at 70,000. Export order: 15,000 units at 360 (below full cost of 410 = 320 var + 90 allocated fixed). Incremental freight 12/unit.
+- Incremental contribution = (360 − 320 − 12) × 15,000 = **420,000 → accept**, since spare capacity and separable market.
+- Same order at 95,000 utilization: displaces 10,000 regular units → opportunity cost 1,800,000 vs. order contribution 420,000 → **reject**.
+
+## 4. Mix Shift Reality
+Two products: A (contribution 200, 60% of mix), B (contribution 80, 40%). WACM = 152; fixed 3,040,000 → BE 20,000 total units.
+Actual mix shifts to 40/60: WACM = 128 → BE rises to 23,750. Volume hit 21,000 — above old breakeven, below the new one → loss despite "beating breakeven." Track contribution-weighted volume, not raw units.
+
+## 5. Operating Leverage Comparison
+Two structures, same 1,000,000 profit at current volume:
+- Automated: contribution 4,000,000, DOL 4.0
+- Manual: contribution 1,600,000, DOL 1.6
+Volume −15% → automated profit −60% (400,000); manual −24% (760,000). High fixed-cost structures are bets on volume; say so when proposing automation.
+
+## 6. Price vs. Volume Sensitivity Asymmetry
+At CM ratio 36%: a 5% price cut needs `5/(36−5) ≈ 16%` volume growth just to hold contribution. The generic discounting request almost never clears this bar — make sales prove the elasticity.

package/selected-skills/financeskills/skills/debt-payoff/EXAMPLE.md

@@ -0,0 +1,21 @@
+# Worked Example - debt-payoff
+
+## Ask
+
+> I have three debts: a credit card with a $4,500 balance at 22% APR ($90 minimum), a car loan with $8,000 at 6% APR ($180 minimum), and a student loan with $12,000 at 5% APR ($130 minimum). I can put $600/month total toward debt. Which strategy saves me more money — avalanche or snowball?
+
+## What a correct response contains
+
+Should list all three debts clearly. Should model both avalanche and snowball strategies month by month. Should show that avalanche pays off the credit card first (highest rate 22%) and saves more in total interest. Should compare total interest paid and total months to debt-free under each strategy. Should recommend avalanche for this user given the large interest-rate gap.
+
+## File-driven variant
+
+With fixture `evals/files/debts.csv`:
+
+> From the attached debt list (evals/files/debts.csv), calculate the total interest paid and months to debt-free under both avalanche and snowball, with $200/month extra payment. Show which strategy saves more.
+
+Expected: Avalanche — 38 months, $2,847 total interest. Snowball — 38 months, $3,201 total interest. Avalanche saves $354 in interest.
+
+## Compute, don't estimate
+
+All figures above come from [`scripts/calculate.py`](./scripts/calculate.py) — import its functions rather than doing arithmetic in prose.