ktx-ai-data-agents

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

• Third-party content exposure detected (medium risk: 0.65). The skill’s runtime context can include wiki content ingested from an outsider-authored Notion workspace (e.g., context_sources: type: notion with recursive: true), where ktx ingest reads Notion page body text and then exposes it to the agent via the MCP tools (e.g., ktx_search_wiki / ktx_get_context_summary).

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill instructs running "npx @kaelio/ktx setup" (which fetches and executes the @kaelio/ktx package from the npm registry, e.g. https://www.npmjs.com/package/@kaelio/ktx) so remote code is retrieved and executed at runtime, presenting a high-confidence runtime external code-execution dependency.