skills.rest
Cogni-AI-OU
Cogni-AI-OUOFFICIAL
Explore all skills by this author ā†’
Websiteš•@Cogni_AI_OUGitHub

critical-thinking

Deconstruct ambiguous technical problems into verifiable success criteria and competing hypotheses.

3|1|Updated Apr 12, 2026
Search Tags:#root cause analysis#hypothesis testing#critical thinking#architecture planning#assumption management#adversarial red-teaming#MiniZinc
One-click install
npx skills add https://github.com/Cogni-AI-OU/cogni-ai-agent-skills --skill critical-thinking-cogni-ai-ou
Or copy as Structured Prompt for Agentā–¼
Please help me install this Agent Skill.
Skill: critical-thinking
Source: https://github.com/Cogni-AI-OU/cogni-ai-agent-skills/tree/main/critical-thinking
Command: npx skills add https://github.com/Cogni-AI-OU/cogni-ai-agent-skills --skill critical-thinking-cogni-ai-ou

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Critical-thinking helps you resolve ambiguous, contradictory, or complex technical problems by separating goals from assumptions, generating alternative hypotheses, and validating plans with systematic, evidence-driven reasoning.

Core Features & Use Cases

  • Deconstruct & Frame: Separates the desired conclusion from premises and explicitly states assumptions and hidden dependencies.
  • Hypothesis Generation & Adversarial Red-Teaming: Surfaces multiple interpretations and attacks the most likely failure points to reduce blind spots.
  • Transform Tasks into Verifiable Goals: Converts vague work into measurable outcomes (e.g., tests to reproduce and prevent bugs, success criteria before changes).
  • Constraint-Aware Planning (MiniZinc when applicable): Uses dry-code constraint modeling to expose restrictions and ambiguity in complex domains.
  • Troubleshooting & Root Cause Isolation: Breaks failures into reproducible steps and narrows faults using logs or version history.

Quick Start

Use the critical-thinking skill to analyze the failing behavior by stating your assumptions, listing competing hypotheses, turning the desired fix into a verifiable test plan, and then stress-testing the proposed approach for likely points of failure.

Frequently Asked Questions about critical-thinking

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

FAQPage Schema
How do I perform root cause analysis for silent failures in my system?ā–¼

Root cause analysis for silent failures requires stating assumptions, listing competing hypotheses, and narrowing faults using logs or version history. This approach transforms vague failures into reproducible steps to isolate the exact fault.

What is the best way to evaluate architecture planning assumptions before implementation?ā–¼

Architecture planning evaluation involves separating desired conclusions from premises and explicitly stating hidden dependencies. Constraint-aware verification, including optional MiniZinc dry-code, exposes restrictions and ambiguity before you commit to changes.

How does adversarial red-teaming improve troubleshooting and refactoring decisions?ā–¼

Adversarial red-teaming improves troubleshooting by attacking the most likely failure points to reduce blind spots. It stress-tests proposed refactoring approaches, ensuring your decisions are validated through evidence-driven reasoning rather than assumptions.

How do I turn ambiguous technical tasks into verifiable success criteria?ā–¼

Turning ambiguous tasks into verifiable goals requires deconstructing the objective and transforming vague work into measurable outcomes. You establish tests to reproduce bugs and define success criteria before making any system changes.

Can I use MiniZinc for constraint modeling when analyzing complex technical problems?ā–¼

Yes, MiniZinc can be used for constraint-aware planning to expose restrictions in complex domains. It generates dry-code constraint models that validate hypotheses and verify assumptions during ambiguous problem analysis.

When should I not use hypothesis testing for troubleshooting system failures?ā–¼

Hypothesis testing is less effective for troubleshooting when you lack access to logs, version history, or reproducible steps. Without evidence to validate or reject competing hypotheses, systematic root cause isolation cannot proceed accurately.

Related Skills

Explore More Software Engineering Skillsā†’
proyecto26proyecto26
42

requirements-scoping

Structure ambiguous system design prompts into functional requirements, non-functional constraints, and out-of-scope boundaries.

Official
Intermediate
Giskard-AIGiskard-AI
5

scenario-generator

Generate adversarial test scenarios and suites for AI agents using giskard.checks.

Official
Advanced
rainoftimerainoftime
17

smt-solver-interface

Encode program properties into SMT formulas and invoke solvers like Z3 or CVC5.

Community
Intermediate
farmountainfarmountain
1

Refactor Agent

Detect code smells and apply incremental refactorings with test coverage.

Community
Advanced