How does this skill handle source credibility?

It systematically evaluates every piece of evidence based on the source's bias, the recency of the data, its direct relevance, and the methodology used to gather it.

What is a 'reference class' in this context?

A reference class identifies similar past decisions and their objective success rates (base rates) to provide a reality check against the 'inside view' of a current project.

What is the main goal of the Deep Research skill?

The main goal is to ground decisions in facts by gathering multi-source evidence and identifying historical reference classes to prevent overconfidence and cognitive bias.

How does it manage different levels of research depth?

It operates on a weighted system (Light, Medium, Complete) that scales research intensity from simple user queries to thorough multi-source academic and market analysis.

Does it prioritize internal or external data?

The skill follows a 'human-first' approach, querying users for internal knowledge and private documents before proceeding to external web research for industry trends.

Deep Research & Evidence Gathering

Name: Deep Research & Evidence Gathering
Author: wme3

bywme3

•

Productivity & Workflow

Conducts comprehensive evidence gathering, reference class analysis, and source evaluation to ground decision-making in verified facts.

Deep Research is a specialized decision-support skill designed to fill knowledge gaps and counter overconfidence bias during the 'Deliberate Decisions' workflow. It systematically investigates internal documents and external sources to gather evidence, identifies reference classes for success-rate benchmarking, and evaluates the credibility of found information. By prioritizing research questions based on impact and actively seeking disconfirming evidence, this skill ensures that critical choices are backed by rigorous data rather than assumptions.

Key Features

013 GitHub stars

02Rigorous source credibility, recency, and methodology evaluation

03Reference class analysis to determine objective base rates of success

04Prioritized research question mapping based on decision impact and feasibility

05Multi-source evidence gathering across internal documents and web searches

06Active bias monitoring to counter confirmation, availability, and authority bias

Use Cases

01Validating technical architecture choices against industry benchmarks and reference classes

02Gathering market intelligence and competitor data to support high-stakes executive decisions

03Conducting due diligence for new product launches or strategic business pivots

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add wme3/deliberatedecisions deep-research

For use in Claude.ai and ChatGPT

Download Skill