skills.rest
markusstrasser
markusstrasserCOMMUNITY
Explore all skills by this author →
GitHub

competing-hypotheses

Generate competing hypotheses and eliminate them to reveal the surviving explanation.

Updated Nov 9, 2025
Search Tags:#analysis#red-team#hypothesis-testing#evidence-evaluation#competing-hypotheses#diagnosticity
One-click install
npx skills add https://github.com/markusstrasser/skills --skill competing-hypotheses-markusstrasser
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: competing-hypotheses
Source: https://github.com/markusstrasser/skills/tree/main/archive/competing-hypotheses
Command: npx skills add https://github.com/markusstrasser/skills --skill competing-hypotheses-markusstrasser

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

ACH helps analysts evaluate claims or anomalies by generating competing hypotheses and eliminating incorrect ones to reveal the surviving explanation.

Core Features & Use Cases

  • Generate at least three competing hypotheses (H1, H2, H3) for the target and add domain-specific sub-hypotheses as needed.
  • Dispatch three parallel agents to evaluate evidence from multiple sources and perspectives, labeling claims with sources.
  • Build a diagnosticity matrix to score evidence across hypotheses and guide synthesis of the surviving explanation.
  • Optional Red Team review to challenge the conclusion and uncover missed evidence.

Quick Start

Provide the target (entity, anomaly, or claim) and run ACH to generate three competing hypotheses, dispatch parallel agents, and synthesize the surviving explanation.

Frequently Asked Questions about competing-hypotheses

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
What is competing hypotheses analysis and how does it evaluate claims?▼

Competing hypotheses analysis evaluates claims by generating multiple plausible explanations, testing evidence against each, and eliminating incorrect hypotheses to reveal the surviving explanation.

How do I evaluate evidence to find the most plausible explanation for an anomaly?▼

You evaluate evidence by building a diagnosticity matrix that scores evidence across competing hypotheses, dispatching parallel agents to test sources, and synthesizing the surviving explanation.

What's the best way to test multiple hypotheses against conflicting data sources?▼

The best way is generating at least three competing hypotheses, dispatching parallel agents to evaluate evidence from diverse sources, and using a diagnosticity matrix to score and synthesize the surviving explanation.

Can I use red team review to challenge a hypothesis conclusion?▼

Yes, an optional Red Team review challenges the final conclusion to uncover missed evidence and test the robustness of the surviving explanation generated by the analysis.

How does the diagnosticity matrix score evidence across different hypotheses?▼

The diagnosticity matrix scores evidence across multiple competing hypotheses by evaluating source-labeled claims, guiding analysts to eliminate weaker explanations and synthesize the final surviving explanation.

Related Skills

Explore More Data & Analytics Skills→
zizhenchen47-osszizhenchen47-oss

legal-abductive-reasoning

Generate and compare competing legal hypotheses from incomplete facts.

Community
Advanced
sorryhyunsorryhyun
12

doc-lookup

Fetch specific documentation sections from DiPeO's docs using anchors or keyword queries.

Community
Intermediate
EAIconsultingEAIconsulting
21

research-synthesizer

Coordinate parallel researchers to synthesize findings with confidence ratings.

Community
Advanced
richardnguyen0715richardnguyen0715

simpo-training

Optimize language model alignment without a reference model using SimPO.

Community
Advanced