local-models

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

• Third-party content exposure detected (medium risk: 0.30). The skill routes user-provided free-text queries (e.g., “ask”, “code”, “deep”, “reason” commands and Claude prompts) directly into local Ollama at runtime via shell/Bash without filtering for trusted authorship, so an outsider can inject arbitrary text through their query.