🌍 OpenClaw Think-Tank Intelligence — 技能工具

技能文档

Produce structured geopolitical, strategic, and policy analysis in a clear think-tank style.

Use this skill to turn complex international, security, and policy questions into decision-useful outputs with explicit assumptions, confidence labels, alternative hypotheses, and practical recommendations.

Best for

• Country and regional risk assessments
• Sanctions/trade exposure analysis for policy or business decisions
• Scenario planning and red-team challenge before high-stakes moves

Not for

• Operational military/security instructions
• Claims that require classified or non-public intelligence
• Deterministic forecasting presented as certainty

60-second preflight

Before deep analysis, confirm:

• scope (topic/theater/time horizon),
• decision audience,
• decision window,
• evidence access status,
• required output mode (brief/report/risk/scenarios/red-team/json).

Ask only blocking questions.

Quick Start

Install:

clawhub install global-think-tank-analyst

Run:

think-tank Analyze US-China tech decoupling risks 2026-2030
think-tank --scenarios Arctic resource competition under climate change 2027-2035
think-tank --red-team Russian hybrid tactics in Eastern Europe

Modes

think-tank [topic]
think-tank --report [topic]
think-tank --risk [topic]
think-tank --scenarios [topic] [timeframe]
think-tank --horizon [topic] [timeframe]
think-tank --red-team [claim or policy]
think-tank --json [topic]

Core Rules

• Separate sourced facts from expert judgment.
• Mark uncertainty explicitly.
• State key assumptions in deep analysis.
• Include at least one alternative hypothesis when ambiguity is high.
• Use a red-team lens to challenge main conclusions.
• Avoid deterministic language in fast-moving environments.
• Recommend expert review for crisis or high-stakes decisions.
• Do not present speculation as fact.

Decision-Grade Additions (standard/deep mode)

• Add numeric ranges for key impact variables (price, growth, inflation, trade, fiscal effects) when relevant.
• Include a compact Evidence Note with 2-6 external sources and timestamp (YYYY-MM-DD), or explicitly mark source access limits.
• Add Go/No-Go (or Trigger/No-Trigger) criteria with thresholds and dates for decision checkpoints.
• End with a 1-2 week validation plan: what to monitor, who should verify, and what would falsify the base case.

Evidence Safety Guardrails (mandatory)

• Never fabricate sources, URLs, dates, or quotes.
• If external evidence access is unavailable, explicitly output EVIDENCE_ACCESS_LIMITED and switch to scenario/hypothesis mode.
• Label key claims as verified, inferred, or unknown.
• Separate facts from inference in the final memo.
• Downgrade confidence when verification is incomplete.

Confidence Labels

• High: well-supported and relatively stable
• Medium: plausible but contested or incomplete
• Low: weakly supported or rapidly changing
• Speculative: forward-looking inference with limited evidence

Framework Selection

Choose the minimum frameworks needed for the task:

• PESTLE: macro context and structural drivers
• Stakeholder analysis: multi-actor dynamics
• Power mapping: leverage and power balance
• Scenario planning: high uncertainty
• SAT methods: ambiguity, bias, politicization
• SWOT: one actor, policy, or institution
• Cross-impact: second-order effects and cascades

Workflow

• Parse the request: topic, theater, horizon, actors, user objective, mode, depth.
• Frame the question: core question, boundaries, decision context, uncertainties.
• Select frameworks: only what is needed.
• Build the analysis: drivers, actors, incentives, constraints, risks, second-order effects.
• Stress-test: assumptions, underweighted actors, breaking triggers, falsification evidence.
• Deliver: findings, risks, options, recommendations, confidence, indicators.

Advanced Playbooks (vNext)

Use these references when quality bar is high or stakes are material:

• Subagent orchestration: references/subagent-orchestration.md
• Confidence scoring rubric: references/confidence-rubric.md
• Regression/eval gate: references/eval-pack.md
• Enterprise architecture: references/enterprise-v1-blueprint.md
• Evidence layer spec: references/evidence-layer-spec.md
• Source policy and provenance: references/source-policy-and-provenance.md
• Governance and audit: references/governance-and-audit.md

Output Formats

Executive Policy Brief

• Executive Summary
• Key Findings
• Main Risks
• Policy or Strategy Options
• Recommendations
• Confidence and Assumptions

Full Strategic Report

• Executive Summary
• Situation Overview
• Context Scan
• Key Actors and Power Map
• Strategic Drivers
• Risk Matrix
• Scenario Analysis
• Alternative Hypotheses
• Policy Options
• Recommendations
• Indicators to Watch
• Confidence and Caveats

Risk Assessment

• Risk Overview
• Risk Matrix
• Trigger Conditions
• Impact Pathways
• Mitigation Options
• Indicators to Watch

Red-Team Memo

• Target Claim or Strategy
• Hidden Assumptions
• Competing Hypotheses
• Failure Modes
• Adversary Perspective
• Revised Assessment

Standard Output Template

# [Title]
Executive Summary
[Concise synthesis]
Situation Overview
[Current context]
Strategic Drivers
Driver 1
Driver 2
Driver 3
Key Actors
| Actor | Interests | Capabilities | Constraints | Likely Behavior |
Risk Matrix
| Risk | Likelihood | Impact | Time Horizon | Notes |
Scenarios
Baseline
Optimistic
Pessimistic
Wildcard
Options
Option A
Option B
Option C
Recommendations
Priority 1
Priority 2
Priority 3
Indicators to Watch
Indicator 1
Indicator 2
Indicator 3
Confidence and Assumptions
Confidence:
Key assumptions:
Alternative hypothesis:

Optional JSON Output

{
  "query": "",
  "mode": "brief",
  "time_horizon": "",
  "summary": "",
  "drivers": [],
  "pestle": {
    "political": "",
    "economic": "",
    "social": "",
    "technological": "",
    "legal": "",
    "environmental": ""
  },
  "stakeholders": [
    {
      "name": "",
      "interests": "",
      "capabilities": "",
      "constraints": "",
      "power": "high",
      "position": "mixed"
    }
  ],
  "risks": [
    {
      "name": "",
      "likelihood": "medium",
      "impact": "high",
      "time_horizon": "",
      "notes": ""
    }
  ],
  "scenarios": [
    {
      "name": "Baseline",
      "description": "",
      "drivers": [],
      "indicators": [],
      "confidence": "medium"
    }
  ],
  "policy_options": [],
  "recommendations": [],
  "assumptions": [],
  "alternative_hypotheses": [],
  "confidence": "medium"
}

Limits

This skill does not:

• replace classified, field, or government intelligence
• guarantee forecasting accuracy
• justify advocacy framed as analysis
• remove the need for expert review in crisis decisions

If evidence is thin, keep output concise rather than padded.