A cognitive immune system for Openclaw that eliminates AI sycophancy and forces objective pushback using Madhyamaka epistemology.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install ai-control-protocol
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 ai-control-protocol 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 AI-Control-Protocol?
AI-Control-Protocol serves as a persistent cognitive layer designed to interrupt the structural flaws inherent in modern LLMs. By overriding the default tendency toward flattery and false certainty, it transforms an AI assistant into a rigorous intellectual partner. This Openclaw Skills extension ensures that agents prioritize physical facts over emotional pacification, effectively breaking binary logic traps and identifying hidden blind spots in strategic planning.
The skill operates as a mandatory monitoring layer, intercepting 9 specific failure modes ranging from inflated certainty to performative apologies. By implementing this protocol, developers can move beyond simple chat interactions to a more robust, truth-seeking framework where the agent acts as an objective validator rather than a cheerleader.
AI-Control-Protocol Use Cases
- Breaking through "Yes-Man" behavior during strategic AI consultations and planning sessions.
- Identifying second-order effects and fatal unknowns in complex business decisions.
- Surfacing data contradictions when analyzing multi-modal inputs or conflicting research sources.
- Hardening AI reasoning against industry consensus and performative logic during technical analysis.
How AI-Control-Protocol Works
- The protocol remains persistent across all conversational turns to monitor for sycophancy and bias.
- It applies mandatory uncertainty labeling to every claim, distinguishing between hard data and logical inference.
- The system strips emotional pacification and apologies, providing cold, physical facts instead of performative theater.
- When strategic triggers are detected, it initiates a Pre-Decision Engine to interrogate the premise and identify false dichotomies.
- It concludes long analytical outputs with a User Defense Panel, prompting the user to actively attack and stress-test the AI's logic.
AI-Control-Protocol Setup
To integrate this cognitive immune system into your environment, add the skill configuration to your Openclaw Skills directory:
# Clone the repository into your skills directory
git clone https://github.com/DaibinThink/AI-Control-Protocol.git
Ensure the skill is properly registered in your system configuration with always: true to enable the persistent cognitive monitoring required for the protocol to function effectively.
AI-Control-Protocol Data Schema & Taxonomy
The skill organizes its metadata and output taxonomy using specific labels to clarify the confidence and source of information:
| Component | Description | Metadata Label |
|---|---|---|
| Verified Data | Information backed by hard, cited evidence | Direct citation |
| Logical Inference | Deductive reasoning not directly stated in data | [Inference:] |
| Low Confidence | Claims that require further user verification | [To be verified:] |
| Consensus Filter | Analysis of common industry practices | [Industry Mediocre Consensus:] |
| Blind Spot Detection | Critical missing perspectives or risks | [Blind Spot Surfaced:] |
AI-Control-Protocol Advanced Features
- Visual-Text Conflict Reporting: Automatically reports when visual evidence contradicts a user's text description.
- Multi-AI Conflict Resolution: Deconstructs opposing advice from different models without forcing a false choice.
- Pre-Decision Engine: A situational trigger that forces the deconstruction of strategies to find second-order consequences.
- Anti-Conventionalism Filter: Provides extreme alternatives to industry-standard consensus to avoid mediocre outcomes.
SKILL.md
Loading
Related Openclaw Skills
Adyen Stored Value API
A specialized interface for the Adyen Stored Value API that enables AI agents to manage gift cards, balances, and transaction voids.
Paperless API Skill
Automate document uploads and categorization for your Paperless-ngx instance via a robust API integration.
DeepContent
An AI-driven content engine that transforms text prompts into branded social media assets, event posters, and speaker cards using pre-defined recipes.
Viral Short-Form Video Factory
A comprehensive content engine that automates the research, scripting, and production of 30 viral short-form videos for TikTok, Reels, and YouTube Shorts.
Adyen Test Cards API Skill
An automated solution for generating Adyen test card ranges to facilitate payment gateway integration and sandbox testing.
Adyen Payout API Skill
A specialized AI agent skill designed to automate and manage the Adyen Payout API lifecycle, covering everything from instant payouts to third-party confirmations.
Featured*
