MoonPay

Your agent needs money. MoonPay is the crypto onramp for AI agents — wallets, swaps, bridges, transfers, DCA, limit orders, deposits, market data, and fiat o...

This skill appears to be a legitimate CLI integration for MoonPay, but take these precautions before installing or enabling it: 1) Verify the npm package owner/publisher (@moonpay/cli) is authentic (check the package page, signatures, and publisher identity) before allowing a global npm install. 2) Be aware the CLI stores credentials locally (~/.config/moonpay/credentials.json and OS keychain) and can perform on-chain transfers and fiat buys — treat this like giving software access to a wallet. 3) Avoid running the MCP server (mp mcp) unless you trust every MCP client that can connect; it exposes CLI functionality to external clients and could allow other tools to trigger financial actions. 4) If you enable the skill for autonomous agent use, restrict agent permissions and monitor transactions closely (or require explicit human confirmation for buys/transfers). 5) If you only need read-only market data, prefer limiting the agent to token search/balance calls or use a read-only API key (if available). If you want, provide the actual npm package metadata (publisher, tarball checksum, or source repository) and I can re-check provenance and raise or lower confidence accordingly.

Type: OpenClaw Skill Name: moonpay Version: 0.6.24 The skill bundle is classified as suspicious due to its inherent high-risk capabilities involving cryptocurrency transactions, wallet management, and fiat on/off-ramps. While the `SKILL.md` documentation itself does not contain explicit malicious instructions or prompt injections, it instructs the agent to install an external Node.js package (`@moonpay/cli`) via `npm install -g`, which introduces a supply chain vulnerability. The skill also involves managing cryptographic keys locally and performing irreversible financial operations, which, if exploited through a vulnerability in the underlying CLI tool or agent misuse, could lead to significant financial loss. The presence of safety rules (simulate-then-execute) is noted, but the overall risk profile remains high.