- Home
- Skills
- Trading & Investing
- Trading Strategy Backtest Auditor
Works with the AI Agents you already use
Trading Strategy Backtest Auditor
Independent quantitative audit of trading backtests to detect bias, overfitting, and unrealistic execution assumptions.
$24.99
Trading Strategy Backtest Auditor
Example session with this skill installed
Audit this strategy. It uses daily BTC/USDT data, entry on RSI < 30, exit on RSI > 70. 0.1% fee. Attached is a CSV ledger with 200 trades and an equity curve showing a 4.0 Profit Factor.
- Read your context and instructions
- Compiled the trading strategy backtest
I have audited the BTC/USDT RSI Mean Reversion strategy based on the provided trade ledger and parameters. While the Profit Factor is high, the audit reveals critical execution and statistical risks.
Executive Audit Summary
- Strategy: BTC/USDT RSI Mean Reversion
- Overall Audit Status: HIGH RISK — Major validation gaps
- Confidence: High (based on full trade ledger)
- Top Finding: Intrabar fill realism and look-ahead bias in exit logic.
Performance Metric Review
| Metric | Reported | Audited | Status |
|---|---|---|---|
| Net Profit | $42,000 | $31,500 | Mismatch (Fee Calc) |
| Profit Factor | 4.00 | 2.85 | Overstated |
| Max Drawdown | 12.0% | 18.5% | Underreported |
| Sharpe Ratio | 2.1 | 1.45 | Adjusted for Vol |
Findings Register
| ID | Severity | Issue | Impact |
|---|---|---|---|
| F-001 | Critical | Look-ahead bias in RSI exit | Exits occurring at daily high instead of close. |
| F-002 | High | Fee omission | 0.1% fee not applied to all exit legs in ledger. |
| F-003 | Medium | Sample size | 200 trades over 4 years is low frequency for crypto. |
Prioritized Remediation Plan
- P0: Re-run backtest using "Next Bar Open" for all entries and exits to eliminate look-ahead bias.
- P1: Apply the 0.1% fee to both entry and exit sides; currently only entry is taxed.
- P2: Perform a parameter sweep on RSI bounds (25/75, 35/65) to check for overfitting.
Next steps
- Provide the source code for the RSI calculation to verify window centering.
- Upload a higher resolution (1h or 15m) dataset to audit intrabar stop-loss behavior.
Connects securely to your tools. The creator never sees your data.
About this skill
Backtests often look profitable because of hidden flaws rather than genuine alpha. This skill performs an independent quantitative audit of trading strategy results to expose data leakage, look-ahead bias, and unrealistic execution assumptions before you risk capital.
What it does
- Bias detection identifies look-ahead bias, survivorship bias, and information leakage in signals.
- Data integrity checks inspects for missing bars, incorrect adjustments, and timestamp alignment issues.
- Execution validation audits slippage models, spread assumptions, and intrabar fill realism.
- Metric reconciliation recalculates Sharpe, Sortino, and drawdown from raw trade ledgers to verify consistency.
- Robustness analysis assesses overfitting risks, parameter sensitivity, and regime concentration.
How it works
- Scope establishment where you provide strategy rules, trade logs, or backtest configuration files.
- Technical audit where the agent inspects the methodology for methodological flaws like future-leaking features.
- Statistical stress-testing to determine if results are driven by a few outliers or genuine edge.
- Verdict delivery providing a prioritized finding register from
CriticaltoInformational.
Frameworks & tools
Works with strategy outputs from VectorBT, Backtrader, Pine Script, Zipline, and custom Python/C++ trade ledgers. Analyzes CSV, JSON, and Parquet data formats.
Why this beats prompting it yourself
Generic prompts often take backtest reports at face value or miss subtle intrabar sequencing errors. This skill follows a structured 10-step quantitative audit workflow that treats every input as untrusted until verified against execution logic.
Use cases
- Auditing a Pine Script strategy before migrating it to a live execution bot.
- Validating a third-party signal provider's historical performance claims.
- Stress-testing a machine learning model for data leakage between training and test sets.
- Identifying why a backtest outperforms live paper trading results.
Known limitations
Cannot verify raw data provenance if only aggregate statistics are provided. Does not execute code or connect to live broker APIs. Audit quality depends on the granularity of supplied trade logs.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Trading & Investing.
Does not execute code or connect to live brokers. Requires granular trade logs for full metric reconciliation. Performance metrics depend on user-supplied cost assumptions.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 17 days ago
Permissions
No special permissions declared or detected