gr-oss-marketing

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

• Third-party content exposure detected (high risk: 1.00). The SKILL.md explicitly instructs the agent to "首次调用时 Claude 应 fetch" files from the public GitHub repo https://github.com/Gingiris/gingiris-opensource (e.g., references/*.md), which are untrusted, user-generated third‑party contents the agent must read and use as deep reference and that can materially influence its actions.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly instructs the agent to fetch files at runtime from https://github.com/Gingiris/gingiris-opensource (references/*.md) to use as "深度参考", meaning remote content would be injected into and directly control the agent's prompts/instructions.