The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest data from external, untrusted sources such as Notion documents, database metadata, and dbt models. This creates a surface area for indirect prompt injection where malicious content within those sources could influence the agent's behavior.
Ingestion points: Ingests wiki content (Notion, markdown files), database metadata (PostgreSQL, Snowflake, etc.), and dbt project definitions as specified in ktx.yaml.
Boundary markers: There are no explicit instructions or delimiters mentioned to isolate ingested external text from the agent's primary instructions.
Capability inventory: The skill provides tools for semantic search across ingested data and enables direct SQL querying capabilities (ktx raw-query, ktx_query).
Sanitization: The instructions do not describe any sanitization or validation of the content fetched from Notion or databases before it is processed by the AI agent.
[EXTERNAL_DOWNLOADS]: The skill requires the installation of external packages and configurations from third-party sources.
Evidence: Recommends global and local installation of @kaelio/ktx via npm and npx.
[COMMAND_EXECUTION]: The skill documentation includes several commands that involve setup and execution of the ktx tool which can perform network and file operations.
Evidence: Commands like ktx ingest and ktx mcp start involve reading from local files, connecting to remote databases/APIs, and starting a local server.

ktx-data-agent-context-layer