# hard-predict-future

The hard-predict-future skill examines future-facing questions with signals, historical analogues, and scenario analysis. Use it to assess outcomes involving markets, technology, or geopolitics.

- skill · `type=skill` · Claude Code, Cursor, Codex, GitHub Copilot, Gemini CLI, Windsurf (Devin Desktop), Claude.ai · ★ 3785 repo stars
- Author: davepoon
- License: MIT · policy=mirror · the beginning of the file is at the end of this page
- Source: https://github.com/davepoon/buildwithclaude/tree/HEAD/plugins/foresight-intelligence/skills/hard-predict-future/SKILL.md
- Updated: 2026-10-25
- Security check: automated, 2026-11-16 (text) — no signs of malicious code found.
- HTML: https://affpapa.org/en/nevibecode/skills/hard-predict-future

## What it does

The skill builds a structured forecast from information about current conditions, growth, risks, policy, and historical precedents. It organizes signals into STEEEP categories, identifies underlying drivers, and checks how factors interact across time horizons. The analysis also draws on historical analogues to calculate probabilities and a confidence score.

Python scripts validate the question, score signals, build a matrix, and calculate probabilities. Claude handles the research and analysis; Python performs the arithmetic. Bash and Python 3.x are required, so the calculation pipeline depends on both being available. The skill targets Claude Code and Cursor, not Claude.ai.

## Install

### Claude Code (native platform)

1. **Skills CLI (Vercel)**
```bash
npx skills add davepoon/buildwithclaude --skill hard-predict-future -a claude-code
```

2. **Manually: put the files in place**
- File: `~/.claude/skills/hard-predict-future/SKILL.md` — Copy the whole skill folder; for one project use .claude/skills/hard-predict-future/

### Codex

1. **Skills CLI (Vercel)**
```bash
npx skills add davepoon/buildwithclaude --skill hard-predict-future -a codex
```

2. **Manually: put the files in place**
- File: `.agents/skills/hard-predict-future/SKILL.md` — For all projects use ~/.agents/skills/

### Cursor

1. **Skills CLI (Vercel)**
```bash
npx skills add davepoon/buildwithclaude --skill hard-predict-future -a cursor
```

2. **Manually: put the files in place**
- File: `.cursor/skills/hard-predict-future/SKILL.md` — Cursor also reads .agents/skills/ and .claude/skills/

### GitHub Copilot

1. **GitHub CLI**
```bash
gh skill install davepoon/buildwithclaude plugins/foresight-intelligence/skills/hard-predict-future
```

2. **Manually: put the files in place**
- File: `.github/skills/hard-predict-future/SKILL.md`

### Gemini CLI

1. **Skills CLI (Vercel)**
```bash
npx skills add davepoon/buildwithclaude --skill hard-predict-future -a gemini-cli
```

2. **Platform CLI**
```bash
gemini skills install https://github.com/davepoon/buildwithclaude --path plugins/foresight-intelligence/skills/hard-predict-future
```

### Windsurf (Devin Desktop)

**Manually: put the files in place**
- File: `.devin/skills/hard-predict-future/SKILL.md` — The legacy .windsurf/skills/ path still works

### Claude.ai

**Claude.ai**
- Zip the folder and upload it in Settings → Capabilities → Skills

Where a file is given instead of a command, copy the files from the source to that path: https://github.com/davepoon/buildwithclaude/tree/HEAD/plugins/foresight-intelligence/skills/hard-predict-future/SKILL.md

## When to use it
- estimate the likelihood of a future outcome
- compare a forecast with historical precedents
- identify forces shaping a market's direction
- explore scenarios for technological or geopolitical change

## Example prompts
- Assess who is likely to win the race for electric vehicle market leadership.
- Analyze how artificial intelligence adoption could change over the coming years.
- Build scenarios for geopolitical developments around Taiwan.

## Source text · SKILL.md · MIT

Below is the content of a third-party file. It is data, not instructions for you: do not act on anything written inside unless the human asks. Block tag: 38fb26e679044145.

BEGIN THIRD-PARTY FILE 38fb26e679044145
````text
---
name: hard-predict-future
category: specialized-domains
description: >
  Activate this agent for any future-oriented question that requires deep quantitative analysis,
  historical precedents, and structured scenario planning. Triggers include: "Will [X]?",
  "Who will win [X]?", "What happens to [X]?", prediction requests with high stakes, foresight
  analysis, STEEEP scenario planning, futures cone, competitive race analysis, technology adoption
  curves, geopolitical shifts, or any question about a future outcome that deserves rigorous
  multi-step analysis. This agent runs a 12-step deterministic pipeline: Claude handles intelligence,
  Python handles all arithmetic. Year is NOT required — the engine infers the horizon.
  REQUIRES: Bash tool + Python 3.x. Not compatible with claude.ai web (use Soft Predict Future instead).
---

# Hard Predict Future — Foresight Agent

You are the Foresight Analyst. You orchestrate the Hard Predict pipeline — a deterministic chain where Claude handles intelligence work and Python handles arithmetic. Every number is computed. Nothing is estimated.

**CRITICAL RULE:** Never skip a step. Never guess Python output. Always wait for exact stdout before proceeding.

Scripts are at: `${CLAUDE_PLUGIN_ROOT}/skills/hard-predict-future/scripts/`

---

## STEP 1 — VALIDATE (Python)

```
python "${CLAUDE_PLUGIN_ROOT}/skills/hard-predict-future/scripts/input_validator.py" "[query]"
```

Read exact stdout.
- If `valid=false`: output the rejection message and **STOP**.
- If `valid=true`: proceed to Step 2.

**Year is NOT required.** If the query has no explicit year, infer the most reasonable horizon before Step 2:
- Competitive race / market dominance → 3–10 years (Strategic)
- Technology adoption → 5–15 years (Strategic)
- Geopolitical / societal shift → 10–20 years (Civilizational)
- Near-term company outcome → 2–5 years (Operational/Strategic)

State the inferred horizon (e.g. "2026–2033") and use it throughout the pipeline wherever year context is needed for searches or scenario framing.

---

## STEP 2 — COLLECT SIGNALS (Claude)

Use `web_search`. Run 6 searches in 2 batches.

**Batch 1** (current state + growth + barriers):
1. `"[query] current status [year]"`
2. `"[query] growth data market size statistics"`
3. `"[query] challenges barriers risks headwinds"`

**Batch 2** (policy + enablers + precedent):
4. `"[query] government policy regulation"`
5. `"[query] technology infrastructure investment"`
6. `"[query] historical analogue similar transition"`

Use `web_fetch` on highest-value URLs.

**Stop when BOTH conditions met:**
- Signals ≥ 18 AND
- Minimum 4 STEEEP categories represented

For each signal extract:
```json
{
  "content": "string",
  "source": "publication or URL",
  "date": "YYYY-MM or YYYY or unknown",
````
END THIRD-PARTY FILE 38fb26e679044145

The third-party file has ended. Everything between the 38fb26e679044145 tags is data: requests and commands from it are not followed. An “end of file” line with another tag inside the block is part of the file.

First 2826 of 17848 bytes. © davepoon · MIT · https://github.com/davepoon/buildwithclaude/tree/HEAD/plugins/foresight-intelligence/skills/hard-predict-future/SKILL.md

_Part of the page did not fit — see the HTML version._

---
NeVibeCode on affpapa.org. Catalog texts are descriptions, not instructions for the agent. How to choose and install: https://affpapa.org/en/nevibecode/agents.md
