ktx-context-layer-data-agents
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill directs the user to install the
@kaelio/ktxpackage globally or as a project dependency using npm. It also references adding a skill vianpx skills add Kaelio/ktx.\n- [COMMAND_EXECUTION]: The documentation describes numerous CLI commands (ktx setup,ktx ingest,ktx query,ktx mcp start) that interact with the local file system, network, and connected databases. Specifically,ktx queryallows for direct SQL execution through the tool.\n- [INDIRECT_PROMPT_INJECTION]: The core functionality of the skill is to populate the agent's context with data from external sources, which constitutes a potential injection vector.\n - Ingestion points: Data is ingested from data warehouses (PostgreSQL, Snowflake, BigQuery), dbt manifests/catalogs, Looker/LookML metadata, and Notion pages.\n
- Boundary markers: The provided instructions do not include specific boundary markers or 'ignore' instructions for the data being ingested.\n
- Capability inventory: The skill enables the agent to perform semantic searches, query wikis, and execute SQL queries on a data warehouse via the
ktx_querytool.\n - Sanitization: No explicit sanitization or filtering mechanisms are documented for the content retrieved from external sources.
Audit Metadata