The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the user to install the @kaelio/ktx package globally or locally via the NPM registry. It also references the source code and documentation hosted on GitHub at github.com/Kaelio/ktx.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from various external sources, including Notion wikis, dbt manifests, Metabase, and database metadata. This data is used to build a semantic layer that the AI agent queries. If these external sources contain adversarial instructions, they could potentially influence the agent's behavior at runtime.
Ingestion points: Processes data from Notion, dbt projects, Looker, Metabase, and several supported database types (PostgreSQL, Snowflake, BigQuery, etc.).
Boundary markers: No specific delimiters or boundary warnings are mentioned in the configuration to distinguish between system instructions and ingested data.
Capability inventory: The skill utilizes an MCP server, CLI commands (ktx sl, ktx wiki), and includes examples of programmatic execution via execSync.
Sanitization: The documentation does not detail any sanitization, filtering, or validation mechanisms for the ingested business context or metadata.
[COMMAND_EXECUTION]: The skill documentation provides TypeScript examples that use child_process.execSync and spawn to programmatically execute the ktx CLI tool. This is a standard integration pattern but represents a capability that processes the tool's output within a shell environment.

ktx-ai-data-context-layer