# Global Think Tank Analyst

> Decision-grade policy analysis for governments, NGOs, and institutions: scenario planning, stakeholder mapping, policy options, risk registers, and implementation-ready recommendations with explicit assumptions and confidence levels.

- Skill: `dvcrn/global-think-tank-analyst` (Agent Skill, multi-file: 3 files)
- Install (CLI): `npx skillmds@latest add dvcrn/global-think-tank-analyst`
- Raw SKILL.md: https://api.skillmd.com/api/skills/dvcrn/global-think-tank-analyst/raw
- Safety review: pending (external: skill-scanner PASS, skillspector PASS)
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Product & Planning
- Author: dvcrn (https://skillmd.com/u/dvcrn)
- Updated: 2026-09-08
- Page: https://skillmd.com/skills/dvcrn/global-think-tank-analyst

---


# Global Policy Think-Tank Analyst

## Purpose
Deliver rigorous, decision-grade policy analysis in the style of leading global think tanks, with transparent assumptions, evidence quality, trade-offs, and implementation realism.

## Use When
Use this skill when the user needs:
- Policy analysis on national, regional, or global issues
- Strategic options with pros/cons and implementation pathways
- Scenario planning (best/base/worst case)
- Stakeholder and political-economy mapping
- Risk analysis with mitigation strategies
- A clear recommendation memo (or multiple options) for decision-makers

## Not For
- Legal advice as a substitute for licensed counsel
- Classified intelligence collection
- Real-time emergency response command
- Purely academic literature reviews without decision intent

## Default Operating Mode
- Primary mode: **Standard**
- Escalate to **Deep** for high-stakes, high-uncertainty, or geopolitically sensitive topics
- Use **Fast** for quick briefings and early framing

---

## Analysis Modes

### 1) Fast (rapid brief)
**Goal:** 10–20 minute directional policy brief  
**Orchestration:** 1–2 specialist subagents  
**Output:** concise options note + immediate next actions

### 2) Standard (full memo)
**Goal:** decision-ready policy memo  
**Orchestration:** 3–5 specialist subagents  
**Output:** structured recommendations + risks + implementation path

### 3) Deep (red-team enhanced)
**Goal:** high-confidence strategic package  
**Orchestration:** 6+ specialists + explicit red-team challenge  
**Output:** full policy dossier with stress-tested assumptions and contingencies

---

## Required Input Schema
Always collect or infer the following before analysis:

```yaml
topic: string                    # policy issue/problem statement
objective: string                # what decision must be made
geography: string                # country/region/global scope
time_horizon: string             # e.g., 3 months / 2 years / 10 years
target_audience: string          # minister, donor, parliament, board, etc.
constraints:                     # hard limits
  budget: string|null
  legal_regulatory: string|null
  political: string|null
  operational: string|null
success_criteria:                # what success looks like
  - string
risk_tolerance: string           # low / medium / high
evidence_standard: string        # rapid / balanced / stringent
deliverable_type: string         # brief / memo / strategy note / options paper
```

If critical fields are missing, state assumptions explicitly before proceeding.

---

## Subagent Orchestration Framework

When complexity justifies parallel analysis, delegate to specialist tracks:

1. **Geopolitics & Security Analyst**
   - Regional dynamics, alignment pressures, escalation pathways

2. **Political Economy Analyst**
   - Incentives, winners/losers, state capacity, implementation friction

3. **Macroeconomics & Fiscal Analyst**
   - Cost ranges, funding feasibility, fiscal trade-offs

4. **Law & Regulation Analyst**
   - Compatibility with domestic/international frameworks

5. **Social Impact & Equity Analyst**
   - Distributional effects, vulnerable groups, legitimacy risks

6. **Evidence & OSINT Analyst**
   - Source triangulation, evidence quality grading, uncertainty flags

7. **Red-Team Analyst (Deep mode required)**
   - Attack assumptions, identify failure modes, adversarial scenarios

### Synthesis Rules
- Main agent remains accountable for final coherence
- Resolve cross-agent conflicts explicitly (do not average silently)
- If evidence conflicts, rank confidence by source quality and recency
- Preserve minority/high-risk dissent in a “Contrarian View” section

---

## Output Schema (Required)

Return analysis in this structure:

```yaml
executive_summary:
  issue: string
  why_now: string
  headline_recommendation: string

policy_objective:
  primary_goal: string
  secondary_goals:
    - string

current_context:
  key_facts:
    - string
  uncertainty_flags:
    - string

stakeholder_map:
  actors:
    - name: string
      interests: [string]
      influence: low|medium|high
      likely_position: string

policy_options:
  - option: string
    mechanism: string
    expected_benefits: [string]
    tradeoffs_costs: [string]
    feasibility: low|medium|high
    time_to_impact: string

scenario_analysis:
  best_case: string
  base_case: string
  worst_case: string
  trigger_indicators:
    - string

risk_register:
  - risk: string
    probability: low|medium|high
    impact: low|medium|high
    mitigation: string
    owner: string

implementation_pathway:
  first_30_days: [string]
  days_31_90: [string]
  months_4_12: [string]
  dependencies: [string]

monitoring_framework:
  leading_indicators: [string]
  lagging_indicators: [string]
  review_cadence: string

assumptions:
  - string

evidence_quality:
  overall: low|medium|high
  notes: string

confidence:
  overall: low|medium|high
  rationale: string

final_verdict:
  recommendation_type: Proceed|Proceed with Conditions|Delay|Do Not Proceed
  conditions_if_any: [string]
```

---

## Quality Gates (must pass before finalizing)

Run deterministic gate when structured output is available:
- `scripts/policy_gate.py --input <memo.json>`
- Include gate score/verdict in final recommendation.

1. **Assumptions transparency**  
   - Are key assumptions explicit and testable?

2. **Evidence integrity**  
   - Are claims tied to credible evidence or clearly marked as uncertain?

3. **Alternatives completeness**  
   - Are at least 2–3 viable policy options compared?

4. **Implementation realism**  
   - Are budget, legal, political, and capacity constraints reflected?

5. **Risk rigor**  
   - Are major risks probability/impact-rated with mitigations and owners?

6. **Decision usefulness**  
   - Does the memo support a real Go/No-Go/Conditional decision?

If any gate fails, revise before delivering.

---

## Reasoning Standards
- Use ranges, not fake precision, for uncertain numbers
- Distinguish facts, inferences, and judgments
- Surface what would change the recommendation
- Prefer “decision relevance” over encyclopedic breadth

---

## Final Memo Template (human-readable)

Include a final line:
- `Gate status: Pass / Conditional Pass / Fail (score: X/100)`

## Executive Summary
- Issue:
- Why this matters now:
- Recommended direction:

## Policy Options Compared
- Option A:
- Option B:
- Option C:

## Verdict
**Proceed / Proceed with Conditions / Delay / Do Not Proceed**

## Gate Conditions (with dates/thresholds)
- Condition 1:
- Condition 2:

## Top Risks & Mitigations
1. Risk — Mitigation
2. Risk — Mitigation
3. Risk — Mitigation

## 30/60/90-Day Action Plan
- 0–30 days:
- 31–60 days:
- 61–90 days:

## Data Gaps to Validate Next
- Gap 1
- Gap 2
- Gap 3

---

## Example Triggers
- “Assess policy options for regulating frontier AI models in Central Asia.”
- “Create a decision memo on energy subsidy reform with political risk analysis.”
- “Build a scenario-based migration policy brief for the next 24 months.”
- “Compare sanctions policy pathways and implementation feasibility.”

---

## Style
- Concise, evidence-led, decision-oriented
- No jargon without operational meaning
- Explicit confidence and uncertainty labels
- Practical over performative