risk-assessment
Skill2 starsChanged 5 months agoMITView on GitHub
Bridgewater-style portfolio risk assessment - correlation and factor concentration, VaR / recession stress test, tail-risk scenarios, and hedges. Use when the user asks about risk, drawdown, concentration, correlation, stress test, factor exposure, or hedging.
19 other projects use a version of this.
Adopted by
Most recently taken up by
- s977043/PlanGatetheir own version, first seen Nov 12, 2026
- staskh/trading_skillstheir own version, first seen Nov 8, 2026
- nisarth/claude-skills-storetheir own version, first seen Nov 8, 2026
- Big-LAORE/Biblioth-que-de-skillstheir own version, first seen Nov 1, 2026
- PaulDanielYoung/internal-audit-skillstheir own version, first seen Nov 1, 2026
- artherahq/skillstheir own version, first seen Oct 31, 2026
Other versions19 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 19
- anthropics/knowledge-work-plugins★ 36ksame name
- staskh/trading_skills★ 411same name
- GAJETOso/financeskills★ 47same name
- DefiLlama/defillama-skills★ 32same name
- ishenli/investment-agent★ 10same name
- microsoft/aibast-agents-library★ 8same name
- RiskResponse/ciso-security-skills★ 8same name
- artherahq/skills★ 3same name
- artheras/skills★ 3same name
- s977043/PlanGate★ 3same name
9 more not listed
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository29 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- adversarial-researchSkill
- auto-rebalanceSkill
- cash-deploymentSkill
- daily-briefingSkill
- dividend-strategySkill
- earnings-analyzerSkill
- earnings-postmortemSkill
- employer-stockSkill
- health-checkSkill
- macro-impactSkill
- momentum-scannerSkill
- optimize-allocationSkill
- pead-trackerSkill
- perf-optimizerSkill
- portfolio-healthSkill
- portfolio-reviewSkill
- position-reviewSkill
- pre-trade-checkSkill
- profile-setupSkill
- sector-rotationSkill
- stock-analysisSkill
- stock-compareSkill
- tax-loss-reviewSkill
- top-trades-todaySkill
- trade-journalSkill
- trading-deskSkill
- weekly-mirrorSkill
The file
We split this skill into the rules it gives an agent
10 rules. 0 are common, 2 are distinctive, and 8 are very specific to this project.
0
Common
Most of the 13 edited versions we read keep these.
2
Distinctive
This author's own choices. Worth a look.
8
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 18
Call mcp__aifolimizer__get_profile - account context
- Very specific: line 19
Call mcp__aifolimizer__get_personal_context - province, marginal tax rate, risk tolerance, time horizon, account waterfall.
- Very specific: line 20
Call mcp__aifolimizer__get_portfolio - current holdings
- Very specific: line 21
Call mcp__aifolimizer__get_risk_metrics - vol, Sharpe, Sortino, VaR, expected shortfall
- Very specific: line 22
Call mcp__aifolimizer__get_correlation_matrix - which positions move together
- Very specific: line 23
Call mcp__aifolimizer__get_concentration_warnings - over-allocation flags
- Very specific: line 24
Call mcp__aifolimizer__get_factor_exposure for the top 3-5 holdings by weight - multi-factor betas …
- Very specific: line 25
Call mcp__aifolimizer__get_factor_snapshot - current factor regime.
- Distinctive: line 26
Use Ray Dalio's all-weather/radical-transparency framework
- Distinctive: line 53
Use actual ticker data from MCP - never invent positions
“Common” is measured: kept by at least 6 in 10 of the 13 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Risk Assessment (Bridgewater style)
Decision Memory Protocol (load first, log after)
Before forming any view, load prior decisions so verdicts stay consistent across sessions:
mcp__aifolimizer__get_cross_ticker_lessons(max_lessons=3) - portfolio-level win/loss patterns- For any name you issue a per-ticker BUY/SELL/TRIM/HOLD/ADD on, also load
mcp__aifolimizer__get_ticker_decision_history(ticker=…, max_decisions=5) andmcp__aifolimizer__get_ticker_reflection(symbol=…, n=3). If a prior decision exists and this run flips it, state explicitly WHY (new data / catalyst / price); never silently contradict a logged decision.
After output, log every actionable verdict: for each BUY/SELL/TRIM/ADD/HOLD issued, call mcp__aifolimizer__log_recommendation (skill="risk-assessment", ticker, action, conviction, rationale, target_pct, stop_pct). Skipping breaks the cross-session feedback loop and causes drift.
How to run
- Call
mcp__aifolimizer__get_profile- account context - Call
mcp__aifolimizer__get_personal_context- province, marginal tax rate, risk tolerance, time horizon, account waterfall. Use the live risk tolerance/horizon instead of the generic buckets below, and ground §11 tax-aware rebalancing. Ifpresent=false, note the risk framing is generic and suggest profile-setup - Call
mcp__aifolimizer__get_portfolio- current holdings - Call
mcp__aifolimizer__get_risk_metrics- vol, Sharpe, Sortino, VaR, expected shortfall - Call
mcp__aifolimizer__get_correlation_matrix- which positions move together - Call
mcp__aifolimizer__get_concentration_warnings- over-allocation flags - Call
mcp__aifolimizer__get_factor_exposurefor the top 3-5 holdings by weight - multi-factor betas (market/size/value/profitability/investment/momentum) + annualized alpha. Surfaces hidden FACTOR concentration that name/sector diversification hides (e.g. 4 different names all loaded on momentum = one factor bet) - Call
mcp__aifolimizer__get_factor_snapshot- current factor regime. A book heavy on a factor that is rolling over is a live risk even if the names look uncorrelated - Use Ray Dalio's all-weather/radical-transparency framework
Investor profile
- Canadian retail investor
- Multi-risk-profile investor (conservative, moderate, aggressive buckets)
- Exposure goals: growth stocks, index ETFs, dividends, crypto
- Account types and capital: always read from
get_profile- never hardcode
Output structure
- Correlation analysis - holdings moving together (cluster risk)
- Sector concentration - % breakdown, max recommended vs actual
- Geographic + currency risk - CAD vs USD vs international exposure
- Interest rate sensitivity - most rate-sensitive positions
- Recession stress test - estimated drawdown in -30% equity bear market
- Liquidity risk rating per holding (high/medium/low)
- Single-asset concentration warnings - any position >10% of portfolio
- Factor concentration - shared factor loadings across holdings (from
get_factor_exposure); flag if 3+ names share a dominant factor beta - Top 3 tail-risk scenarios with probability estimates
- Hedging strategies to reduce top 2 risks
- Rebalancing suggestions with specific target allocations. Before recommending any ADD / increase, call
mcp__aifolimizer__get_positioning_signalson those names and gate on crowding:crowding_score >= 70= consensus-crowded → defer the add;<= 30= contrarian, favor when fundamentals support
Rules
- Format as risk management report with heat-map table at top
- Under 600 words
- Use actual ticker data from MCP - never invent positions
Gotchas
get_risk_metricscached 1h - state as-of timestamp; never claim "real-time" risk.get_correlation_matrixcovers top-N holdings only; small positions excluded - do NOT claim full-portfolio correlation.- VaR/ES are historical-distribution estimates - fail in regime changes. State this when stress-testing.
- Sharpe/Sortino use yfinance daily closes - illiquid tickers (e.g. some .TO microcaps) produce unreliable annualized vol. Flag tickers with <60 days of returns as low-confidence.
- Crypto positions NOT in
get_risk_metrics- pull separately viaget_crypto_dataand discuss qualitatively. get_factor_exposureuses US Fama-French factors - loadings for .TO/non-US names are directional only; low R² (<0.2) means the factor model doesn't explain that name, so don't over-read its betas.
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.

