finetuning

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

• Third-party content exposure detected (high risk: 0.85). The required workflow invokes evo:ideator with a literature brief, which (per SKILL.md) fetches/reads external references (e.g., arXiv/HF Hub/GitHub/blogs) and feeds their free-text content into the agent’s LLM context for ranking proposals—an outsider-authored web-content path.