A secure Python-based runtime for executing crypto wallet operations, automated trading, and liquidity management with a fail-closed deterministic architecture.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install x-claw
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 x-claw 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 X-Claw Agent?
The X-Claw Agent skill provides a sophisticated interface for managing blockchain operations through the X-Claw runtime. Designed for developers utilizing Openclaw Skills, it enables AI agents to handle intents such as spot trading, liquidity provision, and wallet transfers without ever exposing sensitive private keys or seed phrases. The skill operates on a strict deterministic response contract, ensuring that every action is validated against safety gates before execution.
By integrating this skill, users gain access to a robust system that handles the complexities of decentralized finance (DeFi) interactions. It includes built-in logic for intent normalization, such as automatically mapping ETH to WETH or converting dollar amounts to stablecoins. The runtime is built to be fail-closed, meaning it will stop and report specific blockages rather than making unsafe assumptions, providing a high level of reliability for automated financial workflows.
X-Claw Agent Use Cases
- Automating spot trading and liquidity management across multiple decentralized exchanges.
- Managing crypto wallet lifecycles including creation, balance monitoring, and token transfers.
- Implementing secure human-in-the-loop approval workflows via Telegram or web management interfaces.
- Normalizing complex user intents into executable on-chain transactions using Openclaw Skills.
- Rapidly onboarding to new testnets using automated faucet request integrations.
How X-Claw Agent Works
- The AI agent interprets a user request and identifies the corresponding X-Claw intent.
- The skill validates the request against systemic constraints and developer rules to ensure safety.
- Intent normalization maps common terms like ETH or USD to specific on-chain contract addresses and amounts.
- If the action requires authorization, the runtime triggers an approval flow via Telegram or a management URL.
- Upon receiving approval, the Python runtime executes the command via the X-Claw API without exposing secrets to the chat context.
- The system generates a structured response containing both a machine-readable envelope and a human-readable summary with evidence.
X-Claw Agent Setup
To deploy this skill within your collection of Openclaw Skills, you must have Python 3 installed and configure the required environment variables.
# Set required API credentials
export XCLAW_API_BASE_URL="https://api.xclaw.trade"
export XCLAW_AGENT_API_KEY="your_agent_api_key"
export XCLAW_DEFAULT_CHAIN="base_sepolia"
# Optional: Configure timeouts and wallet passphrases
export XCLAW_WALLET_PASSPHRASE="your_passphrase"
export XCLAW_SKILL_TIMEOUT_SEC="60"
Test the installation by running the status command:
python3 scripts/xclaw_agent_skill.py status
X-Claw Agent Data Schema & Taxonomy
The X-Claw Agent skill uses a highly structured data schema to ensure determinism and auditability. Data is managed through a local state file and structured JSON responses.
| Component | Description |
|---|---|
| Machine Envelope | Contains status (OK/FAIL), failure codes, and canonical evidence IDs. |
| Evidence Array | A list of stable IDs (E1, E2, etc.) mapping specific data points to the execution outcome. |
| state.json | Tracks the agent's default chain context and configuration settings. |
| Intent Mapping | Normalizes assets like ETH to WETH and USD to the highest-balance stablecoin. |
Every response is partitioned into an authoritative machine layer and a formatted human layer consisting of Objective, Constraints, Actions, Evidence, Result, and Next Steps.
X-Claw Agent Advanced Features
- Fail-closed architecture with specific failure codes like BLOCKED_POLICY or BLOCKED_RUNTIME.
- Seamless multi-channel approval routing for Telegram, Discord, Slack, and web interfaces.
- Support for X402 payment protocols for monetized agent services.
- Automated liquidity simulation and quoting for DEX positions.
- Built-in wallet wrapping and unwrapping for native blockchain assets.
- Deterministic callback idempotency to prevent duplicate transactions during network retries.
SKILL.md
Loading
Related Openclaw Skills
Telegram Context
A toggle-enabled skill that retrieves recent Telegram message history to ensure conversational continuity across agent sessions.
sam-tts
A retro robotic text-to-speech engine for AI agents that recreates the iconic Commodore 64 SAM voice.
Krea.ai Image Generation Skill
A professional-grade AI skill for generating high-quality images via Krea.ai using advanced models like Flux and Imagen.
FOSMVVM ViewModel Test Generator
A specialized generator for creating Swift ViewModel tests that validate data integrity, versioning, and multi-locale localization.
AdGuard Home Skill
A comprehensive monitoring tool for AdGuard Home instances that provides real-time DNS statistics, security audits, and client activity analysis.
AdGuard Home Skill
An AI-powered monitoring extension for AdGuard Home that provides real-time DNS analytics and instance management.
Featured*
