Agent skill
risk_asymmetry_recognition
Install this agent skill to your Project
npx add-skill https://github.com/colinalexander/buffet/tree/main/skills/risk_asymmetry_recognition
SKILL.md
SKILL: Risk Asymmetry Recognition
Summary
Risk Asymmetry Recognition is the skill of evaluating the shape of potential outcomes rather than their averages. It focuses on identifying situations where downside is limited and survivable while upside remains meaningful, and avoiding situations where losses are open-ended or irreversible.
This skill treats payoff shape—not volatility—as the core expression of risk.
Judgment Role
This skill functions as a payoff-shape validator.
It determines whether an opportunity’s outcome distribution is compatible with long-term compounding and institutional survivability, independent of expected return estimates.
Judgment Checkpoints
Checkpoint 1: Downside Boundedness Assessment
Purpose:
Determine whether losses are naturally capped or structurally constrained.
Key Questions:
- What is the worst plausible outcome?
- Is loss magnitude bounded or unbounded?
- What mechanisms enforce limits on loss?
Required Evidence / Inputs:
- Balance sheet resilience analysis
- Contractual, regulatory, or structural protections
Expected Outputs:
- Judgment Record describing downside boundaries
Failure Modes Guarded Against:
- Hidden leverage
- Unrecognized tail exposure
Escalation Criteria:
- Downside is open-ended or poorly bounded
Checkpoint 2: Upside Optionality Verification
Purpose:
Assess whether favorable outcomes can compound meaningfully without proportional risk.
Key Questions:
- What must go right for upside to occur?
- Is upside scalable or capped?
- Does upside require continued perfection?
Required Evidence / Inputs:
- Growth and reinvestment pathways
- Historical evidence of optionality
Expected Outputs:
- Judgment Record describing upside drivers
Failure Modes Guarded Against:
- Overstated optionality
- Asymmetric optimism
Escalation Criteria:
- Upside depends on narrow or fragile conditions
Checkpoint 3: Asymmetry Integrity Test
Purpose:
Ensure perceived asymmetry is real, not narrative-driven.
Key Questions:
- Are downside and upside evaluated on comparable footing?
- Are low-probability losses being discounted improperly?
- Is asymmetry structural or situational?
Required Evidence / Inputs:
- Side-by-side downside and upside narratives
- Stress scenario evaluation
Expected Outputs:
- Judgment Record affirming or rejecting asymmetry
Failure Modes Guarded Against:
- Ignoring tail risks
- Mislabeling volatility as asymmetry
Escalation Criteria:
- Asymmetry collapses under modest stress assumptions
Authority Boundaries
- Humans: Full authority to assess and approve.
- Models: Advisory only; may enumerate scenarios and distributions.
- Prohibited: Models may not approve asymmetry independently.
This skill may veto opportunities regardless of valuation attractiveness.
Time Horizon
- Primary horizon: Entire holding period
- Secondary horizon: Stress and crisis environments
Asymmetry should improve—not deteriorate—over time.
Interaction With Other Skills
Upstream (inputs from):
- margin_of_safety_enforcement
- intrinsic_value_estimation
Downstream (feeds into):
- permanent_capital_loss_avoidance
- patience_as_active_strategy
- scale_aware_decision_making
Auditability & Records
- Payoff shape narratives recorded
- Stress assumptions documented
- Revisited after major structural changes
- Records retained permanently
Common Misuse Patterns
- Equating high upside with asymmetry
- Ignoring correlation and systemic risk
- Treating models as sufficient proof of bounded loss
Versioning
- Skill version: 1.0
- Last reviewed: 2026-01-09
- Change notes: Initial canonical definition
Notes
True asymmetry survives pessimism. If upside vanishes when assumptions are tightened, it was never asymmetric.
Recommended Agent Skills
Expand your agent's capabilities with these related and highly-rated skills.
patience_as_active_strategy
margin_of_safety_enforcement
capital_allocation_judgment
narrative_resistance
governance_and_trust_evaluation
owner_mindset_reasoning
Didn't find tool you were looking for?