ktx-data-agent-context-layer

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

• Third-party content exposure detected (high risk: 0.75). The skill ingests “company knowledge” from external sources like Notion and Looker (and also wiki content) and then exposes that ingested full-text context to the agent via MCP/CLI, so outsider-authored free text can be read at runtime from those third-party content sources.