Forecasting Expert Knowledge
Teaches an agent structured, calibrated forecasting: base rates, signal weighing, and accuracy tracking.
- Repository
- RightNow-AI/openfang
- Stars
- 18,243
- Licence
- Apache-2.0
- Files
- 2 files
What it does
This skill gives the agent a forecasting playbook based on superforecasting research by Philip Tetlock and the Good Judgment Project. When asked to predict an outcome, the agent starts from a historical base rate, breaks the question into smaller parts, weighs evidence as strong, moderate or weak, and states its answer as a probability with a clear list of assumptions. It also explains how to score past predictions with a Brier score so accuracy can be tracked over time. It is a set of written guidelines and templates, not a tool that fetches live data or makes guaranteed predictions.
Great for
- Analysts or students who want structured, probability-based reasoning about future events
- Teams that make recurring predictions (technology adoption, markets, policy) and want consistent methods
- People who want to learn calibration and how to track forecast accuracy
Skip it if
- Anyone expecting certain or authoritative answers: the skill itself says probabilities are estimates and never 0% or 100%, but outputs can still be wrong and should not drive high-stakes financial, medical or legal decisions alone
- Questions needing up-to-date facts: the skill lists sources to consult (FRED, SEC filings, news) but does not itself retrieve them, so the agent may rely on stale training knowledge unless given data or search tools
Try asking
- “Estimate the probability that a specific product feature ships by the end of next quarter, using base rates and signals”
- “Break down whether a central bank will cut interest rates next meeting into reference-class and evidence steps”
- “Score my last 20 predictions with a Brier score and identify where I was overconfident”
- “Build a reasoning chain with key assumptions for whether a startup will raise a Series B this year”
Install in 3 steps
- 1
Download the skill
The ZIP contains the 2 original files, the Apache-2.0 licence and an attribution note.
predictor.zip - 2
Add it to your agent
Claude Code (all projects)
mkdir -p ~/.claude/skills && unzip predictor.zip -d ~/.claude/skills/Claude Code / Agent SDK (this project only)
mkdir -p .claude/skills && unzip predictor.zip -d .claude/skills/Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.
- 3
Ask for the task
No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.
Get better results
- Give the agent concrete, falsifiable questions with a deadline; vague questions produce vague probabilities.
- Supply your own current data or search results, since the skill does not fetch sources and the agent may otherwise rely on outdated knowledge.
- Keep a log of predictions and outcomes so the Brier score and post-mortem steps have real data to work with.
What's inside
- HAND.toml
- SKILL.md
Redistributed unmodified from RightNow-AI/openfang under the Apache-2.0 licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.