Openclaw X402 Skill

Discover, browse, filter, and pay for x402-compatible API endpoints and MCP tools from the x402 Bazaar — the autonomous discovery layer for agentic payments....

This skill appears to be what it says (discovery + pay-per-call), but the packaging and docs are inconsistent and private-key handling is risky. Before installing or using it: 1) Do NOT paste your primary wallet private key into .env or app configs — create a new, funded ephemeral wallet with only the small amount you are willing to lose. 2) Inspect the actual code repository (git clone URL referenced in README) before pip installing anything; verify requirements.txt and review scripts (agent.py, MCP server code). 3) For MCP use, be aware Node/pnpm are required and the MCP config example stores private keys in desktop config files — avoid putting keys into long-lived config files. 4) Prefer using browse-only mode (no key) for discovery. 5) If you must make paid calls, set conservative MAX_SPEND_PER_CALL and monitor outgoing transactions; consider using a hardware or contract-based account that limits spend. 6) If you want a safer thumbs-up, provide the actual repository code (agent.py and requirements.txt) so it can be audited; without those files the packaging is inconsistent and can't be fully validated.

Type: OpenClaw Skill Name: openclaw-x402-skill Version: 0.1.0 The OpenClaw x402 skill is designed for autonomous discovery and payment for API services using cryptocurrency. It explicitly requires and handles `EVM_PRIVATE_KEY` for making payments, which is a high-risk operation. However, the documentation (`SKILL.md`, `README.md`, `x402-MCP.md`) is highly transparent about this, provides strong security warnings (e.g., 'NEVER commit your private key'), and outlines best practices for safe usage. All external network calls (e.g., to Coinbase CDP, PayAI facilitators, Base RPC) are clearly stated and directly align with the skill's core functionality of API discovery and payment processing. There is no evidence of intentional data exfiltration, unauthorized execution, persistence mechanisms, or prompt injection attempts to subvert the agent's intended behavior for malicious purposes. The skill's actions, while sensitive, are fully aligned with its stated purpose.