Cognitive BiasesPublic knowledge library

Free portable Agent Skill

Decision Making Under Uncertainty

Make a decision without pretending uncertainty has disappeared. Use ranges, reference classes, independent estimates, future-value thinking and explicit review triggers when anchors, recent events or sunk costs may distort judgment.

Agent SkillsHermes AgentOpenClawChatGPT SkillsCodex

Open SKILL.md Back to marketplace

What it does

A repeatable workflow, not another bias-name list.

Use it when

  • A deadline, budget, probability or value is being discussed as one precise number.
  • An early estimate may be anchoring later estimates.
  • A recent success or failure is dominating the next choice.
  • A team is continuing mainly because it has already invested heavily.

It returns

  • Decision and uncertainty frame
  • Range / scenarios
  • Reference-class check
  • Anchor and sunk-cost checks
  • Key assumptions
  • Decision threshold, next step and review trigger

Install

One SKILL.md. No API key.

Hermes Agent

mkdir -p ~/.hermes/skills/decision-making-under-uncertainty && curl -fsSL https://cognitive-biases.github.io/agent-skills/decision-making-under-uncertainty/SKILL.md -o ~/.hermes/skills/decision-making-under-uncertainty/SKILL.md

OpenClaw

mkdir -p ~/.openclaw/workspace/skills/decision-making-under-uncertainty && curl -fsSL https://cognitive-biases.github.io/agent-skills/decision-making-under-uncertainty/SKILL.md -o ~/.openclaw/workspace/skills/decision-making-under-uncertainty/SKILL.md

Other Agent Skills-compatible tools

Download SKILL.md and place it in the skill directory supported by your agent. ChatGPT Skills and Codex support the Agent Skills format; their installation surface can vary by product.

Safety boundary

Useful skepticism without amateur diagnosis.

  • Do not manufacture precise probabilities when the evidence does not support them.
  • Do not treat a range as uncertainty removed; state what remains unknown.
  • Do not count already-spent resources as a reason by itself to spend more.
  • Do not claim a cognitive bias caused the decision.

Instruction-only · no secrets · no executable code · no required network access · CC BY-NC-SA 4.0

Evidence-linked lenses

The skill checks patterns. It does not diagnose people.

Anchoring Effect

Anchoring is a well-supported effect in which an initial numerical value can pull a later estimate toward it. A 2026 meta-analysis covering 2,601 effect sizes found a large overall effect, but also substantial variation across studies. The effect should not be treated as a rule that every number changes every judgment: incidental anchors, anchors from a different dimension, clearly random values, incentives, and some debiasing conditions were associated with smaller or null effects.

Subadditivity Effect

People often give a larger total probability when an event is unpacked into separate possibilities than when the same event is judged as one packed category. The effect is not universal: how the possibilities are described and how typical they are can change or even reverse an unpacking effect.

Availability Heuristic

Availability is a judgment heuristic: people can use how easily examples or scenarios come to mind when estimating frequency or probability. That shortcut is not automatically an error because memorable or accessible examples can correlate with real frequency. Bias appears when accessibility is driven by factors that are not diagnostic of the quantity being judged.

Escalation of Commitment

Sunk-cost effects and escalation of commitment overlap but are not interchangeable. Sunk-cost research asks whether irrecoverable prior investments influence current choices. Escalation of commitment describes persistence or additional resource allocation to a failing course of action and can also be driven by personal responsibility, self-justification, project structure, and other factors.