sglang-minimax-m2-series-optimization

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

• Third-party content exposure detected (high risk: 0.80). The skill's workflow (SKILL.md "Investigation Order" and "Investigation Commands" plus references/pr-history.md) explicitly instructs inspecting still-open upstream GitHub PRs and patches (public URLs like github.com/sgl-project/sglang/pull/...), meaning the agent is expected to read public, user-generated PR content that can materially influence its debugging/porting decisions.