A diagnostic tool designed to identify and resolve conflicting instructions across multiple agent files to prevent AI hallucinations and inconsistent behavior.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install contradiction-detector
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 contradiction-detector 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 Contradiction Detector?
The Contradiction Detector is a vital utility within the Openclaw Skills ecosystem, specifically engineered to maintain consistency across complex multi-agent environments. In many AI agent setups, instructions are distributed across various files like AGENTS.md, SOUL.md, and HEARTBEAT.md. When these files contain conflicting rules, agents may experience silent failures or non-deterministic behavior. This skill systematically scans your workspace to ensure every instruction source is in total alignment.
By cross-referencing twelve distinct contradiction patterns, the tool provides developers with a clear roadmap to fix instruction drift. Whether you are adding a new agent or performing a periodic maintenance sweep, using this skill ensures your AI system remains stable and reliable. It treats instruction clarity as a first-class citizen in the development lifecycle of your Openclaw Skills.
Contradiction Detector Use Cases
- When an agent behaves inconsistently despite appearing to have correct instructions.
- After adding a new agent, cron job, or instruction file to your workspace.
- When a fix applied to one instruction file (like SOUL.md) fails to change agent behavior.
- Performing periodic consistency sweeps to maintain long-term system integrity.
How Contradiction Detector Works
- Discovery: The skill enumerates all instruction sources including workspace files, cron job payloads, and installed Openclaw Skills.
- Inventory Building: It builds a comprehensive inventory per agent to ensure no source is overlooked during the audit.
- Cross-Referencing: The system checks every pair of instruction sources against 12 known contradiction patterns like format duplication and routing mismatches.
- Reporting: It generates a structured report identifying the severity, pattern, and specific evidence of each contradiction found.
- Resolution: Fixes are presented as diffs, which are applied to the workspace only after developer confirmation.
Contradiction Detector Setup
To use this skill, ensure your environment meets the prerequisites and follows the standard installation for Openclaw Skills.
# Ensure the Openclaw CLI is installed
openclaw skill install contradiction-detector
# Verify your environment has the necessary binaries
python3 --version
find --version
cat --version
Contradiction Detector Data Schema & Taxonomy
The Contradiction Detector analyzes a specific set of files and metadata to identify conflicts. It organizes its audit based on the following taxonomy:
| Source Category | Files and Data Scanned |
|---|---|
| Agent Core | AGENTS.md, SOUL.md, IDENTITY.md |
| Operational | HEARTBEAT.md, MEMORY.md, USER.md |
| Automation | Cron job payloads, skill files, hook scripts |
| Configuration | openclaw.json and environment drift |
Contradiction Detector Advanced Features
- Automated diff generation for high-speed instruction alignment and conflict resolution.
- Deep scanning support for 12 distinct contradiction patterns ranging from persona conflicts to schedule mismatches.
- Integration capabilities with HEARTBEAT.md for automated, periodic consistency monitoring of all Openclaw Skills.
- Comprehensive workspace-wide scanning that includes custom hooks and sub-agent payloads.
SKILL.md
Loading
Related Openclaw Skills
Zhang Zara Perspective
A strategic thinking lens for navigating AI-era careers, product development, and personal leverage through a builder-first methodology.
Journal Monitor
An automated research assistant skill that monitors academic journals via ISSN and generates instant bilingual English-Chinese reports.
Pilot Event Replay
A specialized utility for capturing event streams to NDJSON files and replaying them to simulate workloads or debug downstream consumers.
Pilot Event Log
A specialized skill for persistent NDJSON event logging featuring automated rotation, compression, and configurable retention policies.
Code Review
A multi-agent code review system that performs structured analysis on Pull Requests and diffs with high-confidence filtering.
sf-ai-agentforce-testing
A professional testing framework for Agentforce agents featuring dual-track validation, multi-turn conversation checks, and automated coverage analysis.
Featured*
