| name | fetch-yfinance |
| description | One-shot code patterns for downloading price and market data from yfinance inside the E2B sandbox. Load this skill whenever a task requires market or futures data from Yahoo Finance. Load examples.md for working code. |
Fetching market data with yfinance
E2B execution model
Each run_code call is a completely fresh Python process. There is no state,
no variables, and no files from any previous call. Every code block must be
fully self-contained: all imports, all data fetching, and all analysis in one
block.
yfinance is pre-installed in the sandbox. No pip install needed.
What this skill provides
examples.md — Working code patterns for:
- Pattern 1: Single ticker, date range (e.g. WTI crude oil
CL=F)
- Pattern 2: Applying a temporal cutoff for backtesting (do not use data after
as_of)
- Pattern 3: Multiple tickers in one fetch
Workflow
- Call
load_skill_resource("fetch-yfinance", "references/examples.md") to load the patterns.
- Identify which pattern fits your task.
- Combine with other skill examples in the same code block.
Common tickers
| Series | Ticker |
|---|
| WTI crude oil | CL=F |
| Brent crude | BZ=F |
| S&P 500 | ^GSPC |
| Natural gas | NG=F |
| USD/CAD | CAD=X |
Gotchas
ticker.history() returns a timezone-aware DatetimeIndex on recent yfinance
versions. Strip the timezone with .dt.tz_localize(None) after reset_index.- For futures (
CL=F, NG=F), use auto_adjust=False and take the Close
column directly — adjusted close is not meaningful for rolled futures.
- Always sort by date ascending after fetching.
