Credit analysis
AI-powered skills for financial professionals. Comprehensive collection of finance, accounting, audit, and compliance skills for AI agents. IFRS/GAAP compliant with industry-specific applications.
npx -y skills add GAJETOso/financeskills --skill credit-analysisAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 6 stars6 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its author says it does
Copied from the file, not written here
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."
SKILL.md
4.0 KB, as published. Nobody here has run it
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
-
Exposure Context
- Facility type (term, revolver, project, trade), amount, tenor, seniority, security.
- Purpose of funds (the second C — speculative purposes change everything).
-
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
- Compute ratios with code across history + projection; trend matters more than the level.
- 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).
- 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 code-verified math.
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: Deterministic functions for this skill's core computations. Run
python3 scripts/calculate.pyto self-test; import the functions instead of doing mental math.
References
- Covenant Design Guide: 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.