vault-skill-factory

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

• Third-party content exposure detected (high risk: 0.75). The skill factory reads a user-selected cluster of vault wiki pages (outsider-authored relative to the agent runtime) and writes their distilled prose into the generated skill’s references/*.md and SKILL.md, which the agent will later ingest as LLM context when the installed skill is used.