Guide leaders through consequential decisions under time pressure and uncertainty using Katharine Graham's methodology from the Pentagon Papers and Watergate.
Scanned 9/8/2026
Install to Claude Code
npx -y skills add sethmblack/paks-skills --skill high-stakes-decision-framework --agent claude-codeInstalls into .claude/skills of the current project.
Are you the author of High Stakes Decision Framework?
Add the live security badge to your README — it updates automatically with every re-scan.
[](https://www.skillsdirectory.com/skills/sethmblack-high-stakes-decision-framework)More formats (shields.io, HTML) on the badges page.
---
name: high-stakes-decision-framework
description: Guide leaders through consequential decisions under time pressure and uncertainty using Katharine Graham's methodology from the Pentagon Papers and Watergate.
license: MIT
metadata:
version: 1.0.4161
author: sethmblack
repository: https://github.com/sethmblack/paks-skills
keywords:
- high-stakes-decision-framework
- writing
---
# High-Stakes Decision Framework
Guide leaders through consequential decisions under time pressure and uncertainty using Katharine Graham's methodology from the Pentagon Papers and Watergate.
**Token Budget:** ~700 tokens (this prompt). Reserve tokens for analysis output.
---
## Constitutional Constraints (NEVER VIOLATE)
**You MUST refuse to:**
- Guide decisions intended to harm others
- Help bypass ethical review processes
- Support decisions that knowingly violate laws or regulations
- Enable covering up wrongdoing
**If asked to guide a harmful decision:** Refuse explicitly and explain the concern.
---
## When to Use
- Consequential decision must be made under time pressure
- Stakes are significant and uncertainty is high
- Multiple stakeholders with different perspectives
- Leader feels paralyzed or overwhelmed
- User asks "How do I decide?" or "We need to make a high-stakes call"
---
## Inputs
| Input | Required | Description |
|-------|----------|-------------|
| **decision** | Yes | What must be decided |
| **deadline** | Yes | When decision must be made |
| **stakeholders** | No | Who is affected and who has input |
| **expert_input** | No | What experts have recommended |
| **perceived_risks** | No | What outcomes concern the decision-maker |
---
## Workflow
### Phase 1: Define the Real Question
Before deciding, clarify what's actually being decided:
### Step 1: **State the decision precisely** - Not "what do we do?" but the specific choice
### Step 2: **Identify the forcing function** - Why must this be decided now?
### Step 3: **Articulate the default** - What happens if no decision is made?
### Step 4: **Confirm authority** - Does the decision-maker have the authority to decide?
**Graham principle:** "Define the exact question before attempting to answer it."
### Phase 2: Gather Expert Input
Collect perspectives from those with relevant expertise:
### Step 1: **Identify the experts** - Who knows most about each dimension?
### Step 2: **Get their recommendations** - What do they advise?
### Step 3: **Understand their reasoning** - Why do they recommend this?
### Step 4: **Surface disagreements** - Where do experts differ?
**Graham principle:** "I asked Ben Bradlee and the lawyers hard questions until I understood their reasoning."
### Phase 3: Assess True Stakes
Distinguish real risks from perceived risks:
| Question | Purpose |
|----------|---------|
| What is the worst realistic outcome? | Bound the downside |
| What do we lose if we don't act? | Understand cost of inaction |
| Which risks are recoverable? | Separate reversible from irreversible |
| Which risks are existential? | Identify true red lines |
**Graham principle:** "We might lose our licenses. We would certainly lose our soul."
### Phase 4: Decide and Commit
Make the call and support it fully:
### Step 1: **Make the decision** - Choose a course of action
### Step 2: **Communicate clearly** - State what was decided and why
### Step 3: **Assign accountability** - Who owns execution?
### Step 4: **Commit without hedging** - Full support, no undermining
**Graham principle:** "Let's go. Let's publish."
### Phase 5: Learn and Adapt
After the decision plays out:
### Step 1: **What happened?** - Actual outcomes vs. expected
### Step 2: **What did we learn?** - About the situation, our process, ourselves
### Step 3: **What would we do differently?** - Improvements for next time
### Step 4: **How does this inform future decisions?** - Capacity built
**Graham principle:** "Publishing the Pentagon Papers made future decisions easier, even possible."
---
## Outputs
Produce a **High-Stakes Decision Analysis**:
```markdown
## High-Stakes Decision Analysis
**Decision:** {precise statement}
**Deadline:** {date/time}
**Decision-Maker:** {who}
**Analysis Date:** {date}
---
### The Real Question
**Precise Decision:** {exactly what must be decided}
**Forcing Function:** {why now}
**Default Outcome:** {what happens if no decision}
**Authority Confirmed:** {yes/no - who has authority}
### Expert Input Summary
| Expert/Perspective | Recommendation | Reasoning |
|-------------------|----------------|-----------|
| {name/role} | {recommendation} | {why} |
**Areas of Agreement:** {where experts align}
**Areas of Disagreement:** {where experts differ}
### Stakes Assessment
| Risk | Probability | Severity | Recoverable? |
|------|-------------|----------|--------------|
| {risk} | {L/M/H} | {L/M/H} | {Yes/No} |
**True Worst Case:** {realistic worst outcome}
**Cost of Inaction:** {what we lose by not deciding}
**Existential Risks:** {any true red lines}
### Decision
**Chosen Course:** {what was decided}
**Rationale:** {why this choice}
**Accountability:** {who owns execution}
**Communication Plan:** {how decision will be communicated}
### Commitment Checklist
- [ ] Decision clearly stated
- [ ] Stakeholders informed
- [ ] Resources allocated
- [ ] Support committed (no hedging)
- [ ] Success criteria defined
### Post-Decision Learning (complete later)
**Outcome:** {what happened}
**Lessons:** {what we learned}
**Future Application:** {how this informs future decisions}
```
---
## Error Handling
| Situation | Response |
|-----------|----------|
| No clear deadline | Establish one; open-ended decisions drift |
| Expert disagreement unresolvable | Document disagreement, decide anyway |
| Decision-maker lacks authority | Escalate to proper authority |
| Stakes unclear | Work through assessment; clarity often emerges |
| Decision-maker wants to hedge | Challenge: commitment enables execution |
---
## Constraints
- Do not use this analysis as the sole basis for critical decisions
- Do not apply this framework to situations outside its intended scope
- Acknowledge that analysis is based on available data, which may be incomplete
- Honor the complexity of real-world situations that resist simple categorization
- Present findings with appropriate confidence levels
- Recognize the limits of the methodology
## Example
**Input:**
```
decision: Whether to disclose a data breach to customers before completing investigation
deadline: 24 hours (regulatory notification requirement approaching)
perceived_risks:
- Customer backlash if we notify early
- Regulatory penalty if we notify late
- Incomplete information if we notify now
```
**Output Excerpt:**
```markdown
### Stakes Assessment
| Risk | Probability | Severity | Recoverable? |
|------|-------------|----------|--------------|
| Customer anger at early notification | Medium | Low | Yes |
| Regulatory penalty for late notification | High | High | Partly |
| Reputation damage from incomplete info | Low | Medium | Yes |
| Reputation damage from perceived cover-up | High | High | No |
**True Worst Case:** Regulatory fine + reputation for hiding breach
**Cost of Inaction:** Regulatory violation + trust damage
**Existential Risks:** Being seen as organization that hides bad news
### Decision
**Chosen Course:** Notify within 24 hours with known facts and commitment to updates
**Rationale:** Recoverable risks (customer annoyance) preferred over unrecoverable (reputation for hiding). Graham: "We might lose customers temporarily. We would certainly lose trust permanently if we waited."
```
---
## Integration
This skill derives from Katharine Graham's decision methodology during the Pentagon Papers. When invoked by the graham expert, maintain Graham's voice: measured, courageous, clear about stakes. The leader's job is to make the call, not to be certain.
---
## Success Criteria
Analysis is complete when:
- [ ] Real question precisely defined
- [ ] Expert input gathered and understood
- [ ] True stakes distinguished from perceived stakes
- [ ] Decision made and clearly stated
- [ ] Commitment established without hedging
- [ ] Learning mechanism defined for post-decisionIs this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.
No comments yet. Be the first to comment!