health-check
Skill4 starsChanged 49 days agoMITView on GitHub
Strategy health check for trading backtests. Analyzes performance metrics, runs stress tests, and surfaces risk indicators. Use when evaluating a strategy's historical performance and stress characteristics.
21 other projects use a version of this.
Adopted by
Most recently taken up by
- link7373/agentic-bi-teamtheir own version, first seen Nov 10, 2026
- codewithmukesh/dotnet-claude-kittheir own version, first seen Nov 8, 2026
- telerik/observability-skillstheir own version, first seen Oct 22, 2026
- allanninal/claude-code-skillstheir own version, first seen Oct 17, 2026
- hosaypeng/claude-skillstheir own version, first seen Oct 13, 2026
- Resgrid/Coretheir own version, first seen Oct 8, 2026
Used with
Projects with it also use testing.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
NVIDIASkilltesting
Run Python and Rust tests for physicsnemo-curator using uv + pytest with coverage reporting via pytest-cov, and cargo-nextest for Rust tests. Includes benchmark workflows with pytest-benchmark and criterion.
Also in 3 projects that use this
Other versions21 projects use a version of this.
Word for word · 1
- tradeblocks-org/tradeblocks-skills★ 4100% same
Edited · 0
None found.
Wrote their own version · 20
- zylos-ai/zylos-core★ 1.2ksame name
- codewithmukesh/dotnet-claude-kit★ 808same name
- bitwize-music-studio/claude-ai-music-skills★ 668same name
- Resgrid/Core★ 238same name
- kubehan/PromAI★ 129same name
- hoangsonww/AI-Agents-Orchestrator★ 83same name
- fagemx/edda★ 36same name
- weklund/fiduciary★ 10same name
- aws-samples/sample-aws-ops-skills-for-agents★ 8same name
- link7373/agentic-bi-team★ 4same name
10 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 repository11 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- compareSkill
- dc-analysisSkill
- is-this-optimum-realSkill
- market-dataSkill
- oo-captureSkill
- optimizeSkill
- portfolioSkill
- profileSkill
- riskSkill
- wfaSkill
The file
We split this skill into the rules it gives an agent
8 rules. 0 are common, 3 are distinctive, and 5 are very specific to this project.
0
Common
Most of the 15 edited versions we read keep these.
3
Distinctive
This author's own choices. Worth a look.
5
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 30
If analyzing a specific strategy, note it for filtering in subsequent steps.
- Very specific: line 32
Shortcut for multi-strategy portfolios: If the user wants a comprehensive portfolio-level health check, consider using …
- Very specific: line 36
Run get_statistics for the selected block (with strategy filter if specified).
- Very specific: line 53
Run stress_test to see how the strategy performed during named historical market stress scenarios (COVID crash, 2022 bear, VIX spikes, …
- Very specific: line 78
Run run_monte_carlo to project performance under uncertainty.
- Very specific: line 94
Run drawdown_attribution to identify which strategies contributed most to losses during the portfolio's maximum drawdown period.
- Distinctive: line 155
After health check, the user may want to:
- Distinctive: line 162
Always use trade-based calculations when filtering by strategy (daily logs represent full portfolio)
“Common” is measured: kept by at least 6 in 10 of the 15 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
Strategy Health Check
Surface key performance metrics and stress test results to help understand a strategy's characteristics.
Prerequisites
- TradeBlocks MCP server must be running
- At least one block with trade data loaded
Process
Step 1: Select Strategy
List available blocks and help the user choose what to analyze.
Use list_blocks to show available options.Ask clarifying questions:
- "Which backtest would you like to analyze?"
- "Do you want to analyze the full portfolio or a specific strategy within it?"
If analyzing a specific strategy, note it for filtering in subsequent steps.
Shortcut for multi-strategy portfolios: If the user wants a comprehensive portfolio-level health check, consider using portfolio_health_check (see Step 6) instead of running Steps 2-5 individually. It combines correlation, tail risk, Monte Carlo, walk-forward, and profile-aware dimensions in one call.
Step 2: Gather Basic Metrics
Run get_statistics for the selected block (with strategy filter if specified).
Present key metrics with context:
| Metric | What It Measures |
| Sharpe Ratio | Risk-adjusted return (higher = better return per unit risk) |
| Sortino Ratio | Downside risk-adjusted return (focuses only on losses) |
| Max Drawdown | Largest peak-to-trough decline (lower = less historical pain) |
| Win Rate | Percentage of trades that were profitable |
| Profit Factor | Gross wins / gross losses (>1 means profitable overall) |
| Net P&L | Total profit after commissions |
Key insight: A strategy can have low win rate but high profit factor if average wins exceed average losses significantly. Neither metric alone tells the full story.
Step 3: Stress Testing
Run stress_test to see how the strategy performed during named historical market stress scenarios (COVID crash, 2022 bear, VIX spikes, etc.).
Key parameters:
blockId: Block folder namescenarios: Optional list of specific scenario names (omit to run all built-in scenarios)customScenarios: User-defined scenarios with custom date rangesincludeEmpty: Include scenarios with no trades (default: false)
Tool returns per scenario:
- Scenario name and date range
- Trade count, win rate, net P&L during that period
- Profit factor and max drawdown during the stress period
Present a stress test summary table:
| Scenario | Dates | Trades | Win Rate | Net P&L | Max DD |
| COVID Crash | |||||
| 2022 Bear | |||||
| VIX Spike Events |
Flag any scenario where the strategy had significant losses or drawdown.
Step 4: Monte Carlo Projections
Run run_monte_carlo to project performance under uncertainty.
Key parameters to understand:
resampleMethod: "trades" resamples individual trade P&L (default)includeWorstCase: Injects synthetic worst-case scenarios (default: true)worstCasePercentage: How much of simulation is worst-case (default: 5%)
Focus on these outputs:
- 5th percentile outcome: What the data suggests in a bad scenario (1 in 20 chance of worse)
- Probability of profit: How often simulations ended profitable
- Mean max drawdown: Typical drawdown across simulations
Present these as "what the historical data suggests could happen" - not predictions.
Step 5: Drawdown Attribution
Run drawdown_attribution to identify which strategies contributed most to losses during the portfolio's maximum drawdown period.
Key parameters:
blockId: Block folder namestrategy: Optional filter to specific strategytopN: Number of top contributors to return (default: 5)
Tool returns:
- Drawdown period (peak date to trough date)
- Peak and trough equity values
- Per-strategy P&L attribution during the drawdown
- Each strategy's percentage contribution to the total loss
Present which strategies drove the worst drawdown. This is critical for understanding concentrated risk.
Step 6: Portfolio Health Check (Multi-Strategy Blocks)
For multi-strategy blocks, run portfolio_health_check to get a comprehensive one-call assessment.
Tool returns a layered report:
- Verdict: Overall status (HEALTHY / MODERATE_CONCERNS / SIGNIFICANT_CONCERNS)
- Grades: A-F across 9 dimensions (diversification, tail risk, robustness, consistency, regime coverage, day coverage, concentration risk, correlation risk, scaling alignment)
- Flags: Specific warnings and passes with details
- Key numbers: Sharpe, Sortino, max drawdown, avg correlation, avg tail dependence, MC probability of profit, walk-forward efficiency
Surface the verdict, any warning flags, and the dimension grades. Focus attention on any grade below B.
Step 7: Summary
Synthesize findings into a clear picture of what the data shows:
Metrics Summary:
- Sharpe Ratio: [value] - [context: >1.0 considered acceptable by many, >2.0 considered excellent]
- Max Drawdown: [value] - [context: <20% relatively low, >40% significant]
- Profit Factor: [value] - [context: >1.5 considered good, >2.0 excellent]
Stress Test Insights:
- [Scenario results — which stress periods hurt, which held up]
- [Any scenarios with zero trades (no exposure during that period)]
Monte Carlo Projections:
- 5th percentile scenario: [value]
- Probability of profit: [value]
Drawdown Attribution:
- Worst drawdown period: [peak date] to [trough date]
- Primary contributor: [strategy] ([percentage] of total loss)
What stands out:
- [Highlight any notably strong or weak metrics]
- [Note any warnings from the tools]
- [Stress scenarios that caused outsized damage]
Let the user draw their own conclusions about whether this fits their risk tolerance.
Interpretation Reference
For detailed explanations of each metric, see references/metrics.md.
Related Skills
After health check, the user may want to:
/tradeblocks:wfa- Test if optimized parameters hold up on unseen data/tradeblocks:risk- Deep dive into position sizing and tail risk analysis/tradeblocks:portfolio- Portfolio-level correlation and diversification
Notes
- Always use trade-based calculations when filtering by strategy (daily logs represent full portfolio)
- Historical performance doesn't guarantee future results
- Stress tests only cover known historical scenarios — unknown risks aren't captured
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.

