Tearsheet Generator

Generate professional tearsheets with custom SVG visualizations using the QuantStats library. Creates performance reports with MAE analysis, leverage recomme...

Security Scan

The skill generally matches a tearsheet/MAE analysis utility, but it contains several incoherences (hard-coded local dev paths, references to external trading systems and SDKs without declaring credentials or required binaries, and broad allowed-tools) that warrant caution before installation.

Purpose & Capability

The name/description describe tearsheet generation and MAE-based leverage recommendations, which aligns with the included Python helpers. However the SKILL.md repeatedly references external integrations (Nautilus Trader, Hyperliquid SDK, Ray Tune, a local path /Users/DanBot/Desktop/dev/Backtests/backtesting/tearsheets/strategy_comparison_tearsheet.py) and a 'Nautilus Verified' workflow. Those integrations would normally require credentials and/or specific binaries - yet the skill declares no required env vars or binaries. Hard-coded absolute dev paths are unexpected for a distributable skill.

Instruction Scope

Instructions tell the agent to read arbitrary user-specified CSV/JSON files and to import/execute code from an absolute local path; they also instruct running verification with Nautilus Trader and applying leverage via SDKs. The skill's header allows Read/Write/Edit/Bash/Glob which would permit executing shell commands and reading/writing files anywhere the agent can access. There are no explicit limits or safeguards in the instructions about what files/paths or external systems to contact.

Install Mechanism

There is no install spec (instruction-only plus one helper module included). That lowers supply-chain risk. The included Python helper file contains no obfuscated code or obvious network calls. No remote downloads or extraction steps are present.

Credentials

The skill references external services (Nautilus Trader, Hyperliquid) and integration points that normally require API keys, SDK access, or local binaries, but the registry metadata lists no required environment variables, credentials, or config paths. This mismatch means either the integration will silently fail or the runtime may try to access local credentials or prompt the agent to obtain them—both are potential security concerns.

Persistence & Privilege

always:false (good). The skill allows Read/Write/Edit/Bash which lets it create files (tearsheets, JSON) and run commands; that is expected for a report generator but increases risk if the agent is allowed to run autonomously with those tools. The skill does not request persistent platform privileges or alter other skills.

What to consider before installing

Things to check before installing or running: - Inspect or run the code in a sandbox: review the included tearsheet_helpers.py and any referenced script (strategy_comparison_tearsheet) before executing. The helper module looks benign, but you should verify there are no hidden network calls in other files. - Ask the author to explain integrations: the SKILL.md references Nautilus Trader and Hyperliquid SDK but the skill declares no credentials or required binaries. If you expect those features, require the skill to explicitly declare the environment variables and binaries it needs. - Remove or change hard-coded absolute paths: the README points to /Users/DanBot/Desktop/dev/Backtests which is a local dev path — the skill should accept relative or configurable paths instead of hard-coded ones. - Limit agent permissions: because allowed-tools includes Bash/Read/Write/Edit/Glob, run the skill in an isolated environment (container/VM) or reduce allowed tools if possible, especially if you will provide real trading account configs. - Do not run with real account credentials or live trading enabled: any integration that can apply leverage or call exchange SDKs should only be used with test accounts after you confirm what network calls the skill makes. - Request missing details from the publisher: ask for (1) a clear list of external services the skill will call, (2) which env vars/credentials are required and why, and (3) a non-dev installation path or packaging instructions. If you only need local report generation from CSVs and will run everything offline, the skill appears functionally coherent after fixing the absolute paths; if you plan to use verification or SDK features, require the author to make credentials and network calls explicit and auditable.

Tearsheet Generator Skill

About

This skill generates custom tearsheets using the QuantStats library - a Python library for portfolio analytics.

Key Features:

• Custom SVG visualizations (returns, drawdowns, monthly heatmaps)
• Professional HTML tearsheets
• MAE (Maximum Adverse Excursion) analysis
• Leverage recommendations based on risk metrics
• Copyable strategy configurations

Generate comprehensive trading strategy tearsheets with:

• IBM Plex Mono font styling (QuantStats format)
• MAE (Max Adverse Excursion) percentile analysis (p90-p99)
• Optimal leverage recommendations with stop-loss levels
• Fixed Position (Static) vs Full Position (Dynamic) analysis
• 10%, 20%, 30% liquidation buffer calculations
• Full trade list with entry/exit details and MAE stats
• Copyable strategy config text boxes
• Multiple leverage scenario comparisons (1x, 10x, 15x, 20x)

Quick Start

# Generate tearsheet from trades CSV
/generate-tearsheet SOL_MTF_EMA_001 --trades ./trades.csv --capital 10000

# Verify backtest with Nautilus Trader
/verify-backtest SOL_MTF_EMA_001 --trades ./trades.csv

# Test optimal leverage configuration
/verify-mae-lev SOL_MTF_EMA_001 --leverage p95

Commands

/generate-tearsheet

Generate a complete tearsheet with all analysis sections.

/verify-backtest

Verify tearsheet results against Nautilus Trader for accuracy validation.

/verify-mae-lev

Run backtest with optimal leverage config derived from MAE analysis.

Output Files

Each tearsheet generation produces:

• {strategy}_comparison.html - Full HTML tearsheet
• {strategy}_comparison_metrics.json - JSON metrics for programmatic access

Key Sections

1. Key Performance Metrics

• B&H, Fix1x, Dyn1x, Fix10x, Dyn10x columns
• Cumulative Return, CAGR, Sharpe, Sortino, Max DD, Calmar
• Intratrade risk metrics with liquidation distance

2. MAE Analysis & Optimal Leverage

• MAE distribution table (min, mean, p50, p75, p90-p99, max)
• Safe leverage recommendations per percentile
• Stop loss table with % PRICE movement (not position cost)

3. Fixed Position (Static) Analysis

• Leverage table: 5x, 10x, 15x, 20x, 25x, 30x
• Columns: Liq @ %Price, Rec. SL, Max Loss, +10% Buffer, +20% Buffer, Risk Level

4. Full Position (Dynamic) Analysis

• Warning about compounding risk
• Leverage table: 1x, 2x, 3x, 5x, 10x
• Recommendation per leverage level

5. Buffer Analysis Summary

• +10%, +20%, +30% buffers above worst MAE
• Safety check for 10x, 15x, 20x leverage

6. Full Trade List

• All trades with entry/exit times, prices, side, PnL, MAE, MFE, duration
• Scrollable table with sticky headers
• Summary row with averages

7. Strategy Configuration

• Original config text box (copyable JSON)
• MAE-optimized config text box (copyable JSON)
• Backtest methodology description

Dependencies

• Python 3.10+
• pandas, numpy, matplotlib
• StrategyComparisonTearsheet from backtesting.tearsheets

Installation

The skill uses the tearsheet generator at: /Users/DanBot/Desktop/dev/Backtests/backtesting/tearsheets/strategy_comparison_tearsheet.py

Ensure this path is accessible or update the script paths accordingly.