What is an example of 'disconfirming evidence' in tech?

If you believe a new feature is a success, disconfirming evidence would be high churn rates, low session duration, or negative feedback from power users.

Is this skill based on scientific research?

Yes, it incorporates cognitive psychology principles and decision-making frameworks popularized by Nobel Prize-winning psychologists Daniel Kahneman and Amos Tversky.

How does this skill improve my software development process?

It helps you objectively evaluate architectural choices and bug reports instead of assuming your initial solution or preferred technology is inherently correct.

Can I use this for team collaboration?

Absolutely. It provides a shared vocabulary and process for teams to 'disagree and commit' by ensuring all contrary evidence is surfaced before final alignment.

When is the best time to apply the confirmation bias skill?

Use it during high-stakes decision-making, when conducting research, or when you notice you are feeling defensive about a specific conclusion or hypothesis.

Confirmation Bias Mitigation

Name: Confirmation Bias Mitigation
Author: lev-os

bylev-os

0•

生産性とワークフロー

Combats cognitive bias by guiding users to actively seek disconfirming data and treat opinions as testable hypotheses.

This Claude Code skill provides a structured framework for critical thinking and objective decision-making by neutralizing the human tendency to favor information that confirms pre-existing beliefs. Inspired by the research of Daniel Kahneman and Amos Tversky, it guides users through a rigorous process to recognize cognitive filters, reframe opinions as hypotheses, and actively hunt for contradictory evidence. It is particularly valuable during high-stakes software architecture decisions, product strategy evaluations, or complex data analysis where objectivity is essential for success.

主な機能

01Hypothesis reframing tools to turn opinions into testable assumptions

020 GitHub stars

03Step-by-step framework for identifying emotional investment in conclusions

04Detection of common anti-patterns that lead to performative objectivity

05Guided discovery techniques for finding and evaluating disconfirming evidence

06Methods for implementing blinding and pre-commitment in decision-making

ユースケース

01Analyzing product usage data and user feedback to validate market fit

02Evaluating high-cost architectural decisions or technology stack migrations

03Facilitating team discussions to ensure dissenting viewpoints are surfaced and valued

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add lev-os/agents confirmation-bias

For use in Claude.ai and ChatGPT