agentic-forecasting

How to use the search_web tool well when grounding a forecast in recent news — phrase cutoff-aware queries, decide what is worth searching for, and weigh sources. Load this before your first search_web call. No scripts.

How to use the search_web tool well when grounding a forecast in recent news — phrase cutoff-aware queries, decide what is worth searching for, and weigh sources. Load this before your first search_web call. No scripts.

How to use the search_web tool well when grounding a forecast in recent news — phrase cutoff-aware queries, decide what is worth searching for, and weigh sources. Load this before your first search_web call. No scripts.

How to use the search_web tool well when grounding a forecast in recent news — phrase cutoff-aware queries, decide what is worth searching for, and weigh sources. Load this before your first search_web call. No scripts.

Reference guide for the repo concierge catalog — domain filters, the search-then-fetch workflow, and bootcamp routing. Load references/catalog-guide.md before your first answer. No scripts.

How to use the code execution sandbox well — parse the JSON payload (not disk files), compute a couple of useful diagnostics before forecasting, and keep the session stateful within a turn. Load this before writing code. No scripts.

The output contract for producing a structured probability distribution over the rate decision (cut / hold / hike) — the JSON shape, the calibration rules, and how to submit it. Load this ONLY when your task payload asks for a forecast; ignore it for open-ended questions. No scripts.

How to use the code execution sandbox well — parse the JSON payload (not disk files), compute a couple of useful diagnostics before forecasting, and keep the session stateful within a turn. Load this before writing code. No scripts.

The output contract for producing a structured probabilistic forecast — the JSON shape, the calibration and quantile rules, and how to submit it. Load this ONLY when your task payload asks for a forecast; ignore it for open-ended questions. No scripts.

How to use the code execution sandbox well — parse the JSON payload (not disk files), compute a couple of useful diagnostics before forecasting, and keep the session stateful within a turn. Load this before writing code. No scripts.

The output contract for producing a structured probabilistic forecast — the JSON shape, the calibration and quantile rules, and how to submit it. Load this ONLY when your task payload asks for a forecast; ignore it for open-ended questions. No scripts.

How to use the code execution sandbox well — parse the JSON payload (not disk files), compute a couple of useful diagnostics before forecasting, and keep the session stateful within a turn. Load this before writing code. No scripts.

The output contract for producing a structured probabilistic forecast — the JSON shape, the calibration and quantile rules, and how to submit it. Load this ONLY when your task payload asks for a forecast; ignore it for open-ended questions. No scripts.

Governs when and how the adaptive WTI analyst updates its strategy skill. Consult this before calling any strategy mutation tool. The process is deliberately conservative — it resists updating on individual surprises and requires pattern-level evidence before revising strategy.

Code patterns for fitting a linear trend and projecting calibrated forecasts. Always load alongside fetch-yfinance and vol-regime — references/examples.md includes a Full Pipeline Example showing the complete self-contained script from yfinance data fetch through vol regime to final interval output.

Code patterns for classifying the current volatility regime and detecting anomalous recent moves. Always load alongside fetch-yfinance — the examples require a yfinance data fetch at the top of the same script. Load references/examples.md for working code.

The adaptive WTI analyst's current forecasting strategy. Load this at the start of every prediction task. This file is generated — edit the state through the mutation tools, not by hand.

The adaptive WTI analyst's current forecasting strategy. Load this at the start of every prediction task. This file is generated — edit the state through the mutation tools, not by hand.

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.

Diagnostic code patterns for interrogating the WTI price series you have been given — vol regime classification, anomaly detection, and adaptive trend-window selection. Load references/analysis-patterns.md for working code. Load references/wti_benchmarks.json for historical benchmark values to compare against. Run this skill before trend-projection.

Copy-pasteable scikit-learn and numpy code patterns for fitting a linear trend on recent WTI price history, projecting point forecasts to standard horizons, and calibrating 80% prediction interval widths from residual standard errors. Load references/projection-examples.md before writing any trend-projection code.