undocs

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

• Third-party content exposure detected (high risk: 0.70). The skill fetches and displays public third‑party data (e.g., contributors from https://ungh.cc/repos/{owner}/{repo}/contributors and sponsors from a configurable sponsors.api URL) and exposes raw/LLM endpoints and page header tools (Copy Page / Open in ChatGPT) that agents are expected to read, so untrusted user‑generated external content can be ingested into the agent workflow.