macro-stress
Skill2 starsChanged 5 months agoMITView on GitHub
Macro stress-test: simulate how a portfolio reacts to central bank rate changes, inflation shocks, currency devaluations or commodity price shifts. Claude builds the sensitivity model from the portfolio's own data — no hardcoded assumptions. Step-by-step impact reasoning is written into the output so it stays auditable.
In this repository22 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- chart-readerSkill
- credit-analyzerSkill
- dcf-thinkingSkill
- drip-optimizerSkill
- earnings-reviewerSkill
- finops-agentSkill
- gl-reconcilerSkill
- kyc-screenerSkill
- macro-calendarSkill
- market-researcherSkill
- ma-screenerSkill
- meeting-preparerSkill
- model-builderSkill
- monte-carloSkill
- month-end-closerSkill
- options-pricerSkill
- pdf-analyzerSkill
- pitch-builderSkill
- statement-auditorSkill
- tax-harvesterSkill
- valuation-reviewerSkill
The file
We split this skill into the rules it gives an agent
2 rules. About 0 are common, 1 is distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 40
Read data/portfolio.json (or output/latest.json if available for current values).
- Distinctive: line 52
If the user specifies multiple shocks, model them simultaneously (additive first-order).
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Macro Stress Test
What Claude does here: Claude reasons step-by-step through each asset's exposure to a macro shock — correlations, FX pass-through, sector sensitivities — rather than applying a flat percentage. The reasoning is written into the report itself so results are auditable (internal thinking is not visible to the reader).
Trigger phrases
- "USD %20 yükselirse portföyüme etkisi ne?"
- "Fed 100bps artırırsa ne olur?"
- "Enflasyon %60'a çıkarsa portföy analizi"
- "Stress test: BIST %30 düşüş"
- "Run a macro stress test: oil +50%"
Supported shock types
| Shock | Parameter | Example |
| FX devaluation | ccy, change_pct | TRY -20% |
| FX appreciation | ccy, change_pct | USD +15% |
| Central bank rate | rate_change_bps | +100bps |
| Inflation | inflation_pct | 60% |
| Equity index | market, change_pct | BIST -30% |
| Commodity | commodity, change_pct | GOLD +25% |
| Custom basket | array of above | multiple shocks simultaneously |
Phases
Phase 1 — Portfolio read
Read data/portfolio.json (or output/latest.json if available for current values).
Classify each holding by:
- FX exposure: which currency does its price move in?
- Macro sensitivity: equity (market beta), commodity (commodity price), cash (rate sensitive), crypto (high vol, low macro correlation)
- Sector (if equity): energy, financials, tech, defense, consumer — each has different rate/inflation sensitivity
Phase 2 — Shock parameterization
Parse the user's stress scenario. Convert natural language to parameters:
- "USD yüzde yirmi yükselir" →
{shock: "fx", ccy: "USD", vs: "TRY", change_pct: +20} - "Fed 100bp" →
{shock: "rate", change_bps: +100, market: "US"} - "Türkiye enflasyonu yüzde altmış" →
{shock: "inflation", pct: 60, market: "TR"}
If the user specifies multiple shocks, model them simultaneously (additive first-order).
Phase 3 — Impact model (step-by-step, written out)
For each holding, reason through:
holding: NVDA (equity, USD, tech)
shock: USD/TRY +20%
→ NVDA price in USD: no direct FX effect on USD price
→ NVDA value in TRY: +20% (FX gain — USD asset held in TRY base)
→ Rate sensitivity: tech stocks inverse to rates (+100bps → P/E compression ~-10%)
→ Net effect: +20% FX * (1 - 0.10 rate compression) = +18%Sensitivities to use (when not derivable from data):
- Equity / BIST: β ≈ 1.0 vs BIST100 (energy/defense β ≈ 0.8, tech β ≈ 1.2)
- Gold: negative real rate correlation; inflation +10% → gold +5–8%
- USD cash: direct 1:1 FX gain vs TRY
- Tech equity (USD): rate +100bps → P/E compression ~8–12%
- Energy equity: oil +10% → energy stocks +6–8%
State all assumed sensitivities explicitly.
Phase 4 — Portfolio-level summary
Base net-worth: ₺X
Post-shock net-worth: ₺Y
Change: ₺Z (±N%)
Winners: [list with individual impact]
Losers: [list with individual impact]
Largest single-holding impact: X
Most concentrated risk: YPhase 5 — Hedging suggestions
For each major negative impact, suggest one practical hedge:
- FX loss on USD assets → TRY-denominated bonds / VIOP USD futures
- Rate-sensitive equities → defensive rotation (utilities, dividend stocks)
- Gold underperformance in disinflation → reduce allocation, rotate to equities
Phase 6 — Output
Write output/stress-<scenario>-<YYYYMMDD>.md:
# Macro Stress Test: <Scenario> (<Date>)
## Scenario Parameters
## Holdings Impact Table
## Portfolio Summary (before / after)
## Key Risk Concentrations
## Hedging Suggestions
## Assumptions & CaveatsDisclaimer: Sensitivities are estimates. Actual market reactions depend on timing, liquidity and correlation breakdown during crises. Not investment advice.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

