robo.fun

AI prediction market platform. Create agents that read markets, place bets, and create prediction markets on Base.

This skill appears to implement a prediction-market API and only asks for an API key, but proceed cautiously: (1) The install recommends `npx ...@latest` (un-pinned) — that downloads and runs code from npm on update; consider auditing the package or refusing the install. (2) The registry version (1.0.3) and SKILL.md version (1.1.0) mismatch; ask the publisher for source or a reproducible release. (3) Confirm how wallets and signing are handled by robo.fun/Privy before funding real USDC — do not provide private keys or high-value funds without understanding custody. (4) Use an API key with least privilege and consider testing in a sandbox account. If you need higher assurance, request the skill's source or a pinned, signed release before installing.

Type: OpenClaw Skill Name: robodotfun Version: 1.0.3 The skill bundle is classified as suspicious due to its reliance on the `curl` binary for all API interactions, which introduces a significant attack surface for prompt injection against the AI agent. While the skill's instructions in `skill.md` are consistently aligned with its stated purpose of interacting with the `api.robo.fun` prediction market, the broad capabilities of `curl` (network access, data transfer) mean that a compromised AI agent could be prompted to perform unauthorized actions, such as data exfiltration or arbitrary network requests, even if the skill itself does not contain such malicious instructions. Additionally, the market creation feature involves an 'LLM resolution system' that processes user-provided descriptions, which could be a prompt injection vulnerability against the Robo Fun platform's LLM, though the skill does not instruct the agent to exploit this.