A systematic framework for AI agents to evaluate claims using psychological principles and structured critical inquiry.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install critical-thinking-framework
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install critical-thinking-framework using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Critical Thinking Framework?
The Critical Thinking Framework is a robust analytical skill designed for AI agents to scrutinize claims, news, and policies through a rigorous six-step psychological methodology. Inspired by Zimbardo's critical thinking principles, it allows an agent to move beyond surface-level information to uncover hidden biases and logical gaps.
By integrating this skill within the Openclaw Skills ecosystem, developers can ensure their AI assistants provide balanced, evidence-based evaluations rather than echoing unverified information. It transforms a standard LLM into a skeptical, multidimensional analyst capable of verifying source credibility and logical consistency.
Critical Thinking Framework Use Cases
- Fact-checking news reports and scientific claims for potential misinformation or sensationalism.
- Evaluating policy proposals or business strategies by identifying underlying economic and social perspectives.
- Identifying cognitive biases in user inputs or generated content to maintain objective decision-making.
- Verifying the credibility of sources and detecting potential conflicts of interest in media reporting.
How Critical Thinking Framework Works
- The skill begins by verifying the source's authority and checking for any financial, political, or media-driven conflicts of interest.
- It calibrates claim plausibility using the Sagan Standard, ensuring that extraordinary claims are backed by extraordinary evidence.
- The agent categorizes evidence quality, distinguishing between anecdotal personal testimony and repeatable experimental data.
- It scans for cognitive biases such as confirmation bias or emotional triggers that might cloud objective judgment.
- Logic is tested by checking for causal fallacies and identifying potential third-variable explanations for observed phenomena.
- Finally, it integrates biological, social, and economic perspectives to provide a comprehensive, multidimensional assessment.
Critical Thinking Framework Setup
To integrate this framework into your agent, ensure you have the Openclaw Skills environment configured:
# Clone the repository into your skills directory
git clone https://github.com/openclaw/critical-thinking-framework.git
# Add the skill to your agent configuration
openclaw install ./critical-thinking-framework
Critical Thinking Framework Data Schema & Taxonomy
The skill produces a structured JSON output to facilitate programmatic decision-making:
| Component | Description | Metadata Fields |
|---|---|---|
| Analysis Summary | High-level risk and trust assessment | overall_trustworthiness, key_risks, recommended_action |
| Detailed Assessment | Granular breakdown of the investigation | source_credibility, claim_plausibility, evidence_quality |
| Bias & Logic | Mapping of detected logical errors | detected_biases, causal_validity, alternative_explanations |
| Perspective Integration | Multidisciplinary analysis markers | perspectives_considered, complexity_level |
Critical Thinking Framework Advanced Features
- Multi-agent synthesis allowing different Openclaw Skills instances to represent specific perspectives like biological versus social factors.
- Automated risk scoring for rapid rejection, suspension of judgment, or cautious acceptance of information.
- Extensible domain support to tailor the depth of analysis for medical, legal, or psychological contexts.
- Deep cognitive bias detection mapping against known psychological heuristics and emotional triggers.
SKILL.md
Loading
Related Openclaw Skills
Dida365 (TickTick) Task Management
A comprehensive AI skill for managing TickTick/Dida365 tasks, projects, and productivity metrics through an MCP-compatible interface.
Health Goal
A comprehensive tracking system to define, monitor, and achieve SMART health and fitness objectives.
Abstract Summarizer
A professional AI tool designed to condense complex scientific papers into publication-ready structured abstracts while maintaining 100% quantitative accuracy.
Articuler
Articuler is an AI-powered networking assistant that generates personalized cold emails and structured outreach playbooks based on LinkedIn profiles.
Context Optimizer
An aggressive optimization skill designed to minimize input tokens, slash API costs, and prevent context truncation in AI agent sessions.
Doctor Prep
A specialized preparation assistant that organizes symptoms, medication lists, and medical history into a professional checklist for doctor appointments.
Featured*
