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). Outsider-authored free text can be ingested at runtime from configured external sources (e.g., Notion page bodies, wiki markdown files, dbt/Looker/Metabase metadata) and then fed into the agent via the MCP tools’ semantic search/context results; this includes Notion/Looker/Metabase content authored by non-operating users.

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

• Potentially malicious external URL detected (high risk: 0.90). The README instructs runtime installation/usage that fetches and executes remote packages (e.g., "npm install -g @kaelio/ktx" and "npx skills add Kaelio/ktx" which pull code from the npm/GitHub package @kaelio/ktx / Kaelio/ktx), so external code is fetched and executed and thus can directly control the agent's behavior.