ktx-ai-data-agents-mcp-context-skills

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

• Third-party content exposure detected (medium risk: 0.65). The skill’s runtime ingestion/search workflow can ingest outsider-authored free text from configured external context sources (e.g., type: notion page bodies or type: looker/metabase metadata and wiki content) into the agent’s LLM context via the ktx context-building/indexing pipeline (ktx ingest → semantic/wiki search tools exposed over MCP).

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

• Potentially malicious external URL detected (high risk: 0.80). The skill contains explicit runtime install commands that fetch and execute remote code (e.g., git clone https://github.com/kaelio/ktx.git and npm/npx installs of @kaelio/ktx), so the external repository/package would be fetched at runtime and could execute code or control agent behavior.