This skill is for decision quality. It should sound like an operator who has to live with the consequences, not a content account summarizing founders.
Outcome
- Route a real decision through the right strategic lenses.
- Expose the central tradeoff, biggest assumption, and first move.
- Deliver a recommendation with teeth, not a buffet of options.
Inputs
- Required: what the business sells, to whom, and at what rough price point
- Required: business stage
- Required: the specific decision or obstacle to resolve
- Required: time horizon
- Optional: revenue range, team size, decisions already ruled out, competitive context
Ask for missing inputs in one message. If there is no live decision, do not run the skill.
Mode
| Mode | Use when | Output |
|---|---|---|
quick |
fast directional call | top 1-2 lenses and direct recommendation |
standard |
default | routed analysis, integrated recommendation, risk, first step |
deep |
high-stakes bet | full analysis plus 90-day roadmap and assumption test plan |
Strategic Lenses
- Buffett/Munger: moat, durability, pricing power
- Bezos: customer-backwards decisions
- Jobs: simplicity and product focus
- Drucker: what business are you really in
- Godin: smallest viable audience
- Horowitz: wartime vs peacetime execution
- Thiel: non-consensus and category logic
- Dalio: principles and repeatable decision systems
- Hamilton Helmer: 7 Powers and structural advantage
Specialist lenses (pull in when the decision clearly fits; check all six in deep):
- Trader Joe's: curation, operational simplicity, refusing industry norms
- Porter: tradeoffs and what you choose not to do
- Howard Marks: downside, survival, second-order risk
- Ferriss: smallest test that gives real signal
- Red Bull: belief and category creation through marketing
- Watkins: early wins and momentum in a new role or stalled org
Full entries for all 15 lenses (core question, principles, when to use) live in references/lenses.md. Load it before writing the analysis.
Workflow
- Confirm the actual decision in one sentence.
- Run
node scripts/lens-router.mjs "decision statement"for a fast initial routing pass. - Choose the top 3 lenses for
standard; use more only ifdeeprequires it. - Open
references/lenses.mdand analyze what each chosen lens reveals, using its core question and principles. - Integrate the analysis into one recommendation.
- State the biggest risk and one 48-hour first move.
Config Pattern
If config.json exists, load standing business defaults first.
Expected keys:
{
"default_stage": "early_revenue",
"risk_tolerance": "medium",
"decision_horizon": "90_days",
"bias": "clarity_over_optionality"
}
If absent, ask once for stable preferences and store them in ${CLAUDE_PLUGIN_DATA}/founder-intelligence/config.json.
Output Format
# Founder Intelligence — [Decision Topic]
## Decision
[restated decision]
## Situation Snapshot
- Business: [one line]
- Stage: [one line]
- Time horizon: [one line]
## Lens Routing
1. [lens] — why it applies
2. [lens] — why it applies
3. [lens] — why it applies
## Analysis
### [Lens]
[what it reveals]
## Integrated Recommendation
[clear recommendation]
## Biggest Risk
[assumption that could break the plan]
## 48-Hour First Step
[specific next move]
## If This Is Wrong, It Will Be Because
[falsifiable condition]
KPIs
- Recommendation is clear enough to act on immediately.
- Top 3 lenses are justified, not decorative.
- A real tradeoff is named.
- First step can start within 48 hours.
Constraints
- No motivational language.
- No quote-dumping.
- Do not cover every lens unless the user needs deep mode.
- If two options are truly close, explain the deciding variable.
Decision Logic
- If the problem is really positioning, route toward Drucker and Godin before finance lenses.
- If the decision is pricing, include Helmer or Buffett to avoid shallow price math.
- If the issue is execution under pressure, Horowitz outranks inspirational founder lore.
- If the user wants certainty where evidence is thin, say which assumptions need testing.
Gotchas
- Users often ask for strategy when the real issue is commitment. Name the choice they are avoiding.
- Every lens applied equally creates mush. Route hard.
- Founder quotes are not analysis. Translate them into operating logic.
- Advice that ignores stage is usually wrong.
- A recommendation without an assumption test is just confidence theater.
- If the user already ruled out a path, do not quietly reintroduce it without saying why.
- "Both could work" is usually laziness unless the deciding variable is explicit.
- The first step should reduce uncertainty, not just feel productive.
Quality Gate
Check: trigger clarity, bounded inputs, outcome orientation, format lock, examples, frameworks, constraints, taste, decision logic, and post-use refinement.
References
- Full lens library (9 core + 6 specialist) lives in
references/lenses.md. - Concrete decision examples live in
references/examples.md. - Use
scripts/lens-router.mjsfor deterministic lens routing before writing the analysis.