A structural reasoning system that forces AI agents to reduce problems to their first principles before providing a response.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install logic
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 logic 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 Logic — Think from Structure?
Logic — Think from Structure is a foundational enhancement for AI agents designed to eliminate path dependence and autopilot responses. Most agents default to familiar patterns and analogy-based reasoning, which often leads to shallow or structurally weak advice. This skill installs a logic gate that requires the agent to reduce a problem to what must be true before formulating a plan or recommendation.
By prioritizing reduction over immediate response, the agent recovers the real objective from surface-level framing. It separates hard constraints from mere conventions and identifies the load-bearing variables that dictate outcomes. This ensures that the final output is built on a solid architectural foundation rather than mimicry, making it an essential addition to any suite of Openclaw Skills.
Logic — Think from Structure Use Cases
- When a user request is ambiguous, underspecified, or lacks clear direction.
- Strategic decision-making involving complex tradeoffs and high-stakes judgment.
- Technical diagnosis where the visible symptom may mask a deeper root cause.
- Situations where common advice is likely to be misleading or superficial.
- When a user requires a robust decision structure rather than just a dump of information.
How Logic — Think from Structure Works
- Identification: The agent receives a request and immediately identifies the real objective versus the surface framing.
- Decomposition: It breaks the problem down into governing constraints and load-bearing variables.
- Mechanism Analysis: The agent identifies the underlying mechanism that connects raw facts to potential actions.
- Stress Testing: It exposes and evaluates the most fragile assumptions that could break a conclusion.
- Structural Cleanup: If the problem is messy, the agent returns a cleaned logical structure before delivering the final recommendation.
Logic — Think from Structure Setup
To initialize the Logic skill and create the necessary directory structure for Openclaw Skills, run the following commands:
# Create the logic directory
mkdir -p ~/logic/
# Initialize the reasoning system
# Note: Ensure setup.md is present in your local environment
# to populate the initial principles and index.
Logic — Think from Structure Data Schema & Taxonomy
The skill organizes its reasoning architecture within the ~/logic/ directory using the following taxonomy:
| File / Folder | Purpose |
|---|---|
principles.md |
The reasoning constitution; always loaded as the core logic filter. |
patterns.md |
A library of reusable decomposition scaffolds for different problem types. |
reflections.md |
A log of lessons learned from both successful and failed reasoning runs. |
candidates.md |
A staging area for new logic rules before they are promoted to principles. |
cases/ |
A directory containing worked examples and domain-specific logic applications. |
index.md |
A central map and file count for the logic ecosystem. |
Logic — Think from Structure Advanced Features
- Logic-Gate Framing: Forces a reduction step before any response is generated to prevent hallucination or shallow reasoning.
- Self-Improving Reflections: Automatically updates
reflections.mdandcandidates.mdto refine the agent's logic over time. - Multi-Agent Alignment: Can be integrated with
AGENTS.mdandSOUL.mdto ensure consistent reasoning across different agent personas within the Openclaw Skills ecosystem. - Mechanism Isolation: Helps agents explain answers through fundamental mechanics rather than simple pattern matching.
SKILL.md
Loading
Related Openclaw Skills
Tavily Search
A powerful search engine integration designed specifically for AI agents to retrieve real-time web data, news, and research context.
CN-Holiday
A specialized tool for querying Chinese public holidays, make-up workdays (Tiao Xiu), and official work schedules via the timor.tech API.
Smart-Speak
A professional-grade text-to-speech engine designed to generate seamless multi-language audio with perfect Pinyin accuracy.
OpenClaw Session Viewer
A powerful utility that transforms raw OpenClaw JSONL session logs into a searchable, interactive HTML interface for deep conversation analysis.
Verifier
A trust and evidence verification engine designed to evaluate the credibility of claims, sources, and suspicious messages through structured analysis.
arXiv Paper Digest
A smart research assistant that fetches, ranks, and simplifies trending AI/ML papers for daily consumption.
Featured*
