Sdk

AI-powered sports betting simulations with Monte Carlo analysis

This package appears to be a legitimate SDK for paid sports-simulation calls that uses Solana payments, and the code shows local transaction signing and API polling — which is expected. However: 1) Do NOT provide your main Solana private key to the agent. A private key in an environment variable gives full control of that wallet (ability to transfer funds). 2) Verify the package source and integrity (npm page, repository, checksums). The skill metadata here shows 'Source: unknown' even though package.json references edgebets.fun and a GitHub URL — confirm those links and check the upstream repo history and npm publisher. 3) Test with a throwaway wallet containing minimal funds (e.g., only enough USDC/SOL to run a single simulation) before using any real funds. 4) Prefer using a signing wallet adapter (wallet adapter or hardware wallet) that never exposes the raw secret to the runtime, or create a dedicated wallet with limited balance. 5) Verify the TREASURY_WALLET address and endpoints (api.edgebets.fun) independently. 6) If you need higher assurance, review the full payment-related code paths (payment.ts/simulation.ts) to confirm no network calls leak secret material and to confirm signature/proof flow is local. If you cannot verify the source, treat this skill as risky for holding any non-trivial funds.

Type: OpenClaw Skill Name: sdk Version: 1.0.2 The EdgeBets skill bundle provides a legitimate sports betting simulation service using the x402 payment protocol on the Solana blockchain. It requires a SOLANA_PRIVATE_KEY to sign 0.50 USDC transfers for paid simulations, but the SDK (src/services/payment.ts) performs all signing locally using standard libraries and only transmits the transaction signature to the API (https://api.edgebets.fun) as proof of payment. The instructions in SKILL.md are transparent regarding costs and functionality, and no evidence of credential theft, unauthorized data exfiltration, or malicious prompt injection was found.