Works with the AI Agents you already use
Deptho
Replay crypto perp backtests against historical trade prints to audit slippage, funding, and fee drag.
Free
Deptho
Example session with this skill installed
I have a BTCUSDT perp backtest. Replay these orders against the aggTrades file and give me a tearsheet with 5bps fees and the calculated slippage. BTC price is 65000.
- Read your context and instructions
- Compiled the deptho
- Generated the document
I have replayed your 3 market orders against the provided historical trade data. The simulation shows a mean slippage of 2.45 bps. I have calculated the net PnL by accounting for this slippage, a 5.0 bps flat fee, and the funding rates found in the data.
Backtest Audit Results
| Metric | Value |
|---|---|
| Total Orders | 3 |
| Fill Rate | 100% (3/3) |
| Mean Slippage | 2.45 bps |
| Median Slippage | 1.80 bps |
| P90 Slippage | 4.10 bps |
| Total Fees | -15.00 USDT |
| Total Funding | -2.10 USDT |
| Total Slippage Cost | -7.35 USDT |
| Net PnL | +145.55 USDT |
Performance Breakdown
Gross PnL: 170.00
Fees: -15.00
Funding: -2.10
Slippage: -7.35
-------------------
Net PnL: 145.55
Next steps
- Review the
report.htmlfile for the full interactive tearsheet and equity curve. - Check the
fill_gap_card.svgto see the visual spread between your backtest and reality. - If the fill rate drops, try increasing the
max_waitparameter in the replay settings.
report.html
PDF · document
Example file from a real run - the skill writes it into your workspace.
Connects securely to your tools. The creator never sees your data.
About this skill
Backtests often report unrealistic returns because they ignore execution slippage, funding rates, and tiered fees. This tool replays market orders against historical trade prints to calculate actual execution costs and net PnL.
What it does
- Simulate fills by replaying market orders against historical trade data using a VWAP-based trade print model.
- Calculate slippage by comparing the replayed fill prices to the reference price at the time of the order.
- Analyze cost drag by decomposing strategy returns into gross PnL, trading fees, funding payments, and slippage.
- Generate tearsheets in a self-contained HTML format containing performance metrics and fill statistics.
- Visualize gaps between theoretical and realistic returns using shareable fill gap cards.
How it works
- Load data: Import your strategy's market orders and historical trade data from Binance archives or Tardis.dev CSVs.
- Replay orders: Run the fill simulation to find the realized price for every trade in the sequence.
- Audit costs: Calculate the mean slippage in basis points and apply it alongside funding and flat fees to your positions.
- Export reports: Generate a full HTML tearsheet or a fill gap card to document the impact of market friction.
Frameworks & tools
Built for Python environments using pandas and numpy. It integrates via a standalone library or through deptho-mcp for LLM-native workflows. Supports Binance archive formats (aggTrades, klines, fundingRate) and Tardis trades.
Why this beats prompting it yourself
Writing reliable fill simulators requires complex logic to handle base unit quantities and time-series alignment without lookahead bias. This skill provides a tested execution model that accurately handles basis point calculations and funding-time accounting.
Use cases
- Auditing a crypto perpetual strategy to see if slippage kills the edge.
- Creating an investor-ready report that shows net returns after all friction costs.
- Identifying how many orders fail to fill within a specific time window.
- Splitting performance attribution into market alpha vs. execution cost.
Known limitations
Does not model order book depth or queue position; orders are replayed against the tape without market impact. Requires local CSV or ZIP files for trades and bars. Only flat fee rates are supported.
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.
Uses trade prints, not order book depth. No market impact modeling. Requires local data files. Mean slippage < 0 is treated as 0 cost. Supports flat fees only.
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
Free to use
No account required to browse
Trust & safety
Security scan
Security scan passed 10 days ago
- Free with an Agensi account