grok

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 0.80). The set includes a direct installer script (https://x.ai/cli/install.sh) intended to be piped to a shell—which is a high-risk pattern even if x.ai appears official; docs.x.ai is benign documentation and github.com/user/repo.git is a generic/ambiguous repo reference, so overall the presence of the remote install.sh makes this suspicious.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.75). Outsider free text can enter the Grok LLM context via the “Clone-to-temp Review” runtime path: the skill runs git clone https://github.com/user/repo.git ... && gh pr checkout 42 && grok ..., and Grok will read the PR’s files/README/comments from that outsider-authored GitHub repository as readable text during the review.