The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs users to install the @kaelio/ktx package from the NPM registry and references the Kaelio/ktx repository on GitHub for skill registration.
[COMMAND_EXECUTION]: To function, the skill requires the execution of multiple CLI commands including ktx setup, ktx ingest, and ktx mcp start, which interact with local files and remote databases.
[INDIRECT_PROMPT_INJECTION]: The skill acts as a context layer that ingests data from external sources such as data warehouses (PostgreSQL, Snowflake, BigQuery), Notion documents, and dbt metadata.
Ingestion points: Reads metadata and content from databases, Notion APIs, and local Markdown/YAML files via ktx ingest.
Boundary markers: Not explicitly defined in the provided configuration; relies on the underlying LLM's handling of search results.
Capability inventory: The agent can search the semantic layer, preview SQL queries, and search wiki content using the provided MCP tools.
Sanitization: No specific sanitization or filtering of ingested data is mentioned in the documentation.
[CREDENTIALS_UNSAFE]: The documentation provides examples for setting environment variables like ANTHROPIC_API_KEY, DB_PASSWORD, and NOTION_TOKEN. However, it follows security best practices by advising that these secrets be stored in a .ktx/secrets.yaml file that is explicitly excluded from version control.

ktx-ai-data-agents-context