How does this skill improve technical post-mortems?

It provides a step-by-step process to reconstruct what was known during the incident, helping teams identify actual architectural gaps rather than just dismissing the event as a 'silly mistake'.

Why should I evaluate decisions independently of their outcomes?

A good decision can have a bad outcome due to luck, and a bad decision can have a good outcome. Evaluating the process ensures you repeat good logic and fix bad reasoning, regardless of the immediate result.

Can this skill be used for strategic planning?

Yes. By reviewing past predictions and documenting current assumptions prospectively, you build a more accurate model of your team's predictive capabilities and avoid overconfidence.

What is hindsight bias in a software development context?

It is the tendency to believe a bug or system failure was 'obvious' only after it happened, which often leads to unfair blame and a failure to address the actual systemic risks that were invisible at the time.

Hindsight Bias Analysis

Name: Hindsight Bias Analysis
Author: lev-os

bylev-os

0•

生産性とワークフロー

Mitigates the 'knew-it-all-along' effect during project retrospectives to improve decision quality and future predictability.

The Hindsight Bias skill provides a structured framework to help teams and individuals objectively analyze past events without the distorting influence of known outcomes. By actively separating 'before' knowledge from 'after' results, it enables users to conduct more accurate post-mortems, identify genuine signals amidst noise, and evaluate decision-making processes based on the information available at the time. This skill is vital for fostering a blameless culture, improving strategic planning, and preventing the overconfidence that often follows accidental successes or predictable failures.

主な機能

01Temporal knowledge mapping to separate pre-event data from outcomes

020 GitHub stars

03Counterfactual thinking exercises to simulate alternative results

04Outcome-blind decision analysis for objective process evaluation

05Integration of contemporaneous records like decision logs and memos

06Signal vs. Noise differentiation for retrospective data calibration

ユースケース

01Evaluating hiring and performance decisions fairly without outcome bias

02Conducting engineering post-mortems after system outages or incidents

03Reviewing product strategy and market pivots to identify actual learnings

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add lev-os/agents hindsight-bias

For use in Claude.ai and ChatGPT

Download Skill