external-gitcode-ascend-ascend-inference-repos-copilot

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 query and ingest content from public open-source repositories via the deepwiki MCP tools (e.g., repoName mappings such as vllm-project/vllm and verylucky01/MindIE-*) so untrusted, user-generated third-party docs/code are read and used to drive responses and follow-up decisions.