The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill's primary purpose is to generate TypeScript code and JSON configuration files for 'Flow' and 'OData' datasources. It uses the dxs datasource generate-flow and dxs datasource generate commands to create executable logic for the Datex Studio platform.
[INDIRECT_PROMPT_INJECTION]: The skill defines a workflow where it ingests data from external sources such as requirements briefs and schema metadata.
Ingestion points: Requirements briefs (provided by other skills or users) and metadata retrieved via schema-explorer are used to drive the generation of datasource fields and logic.
Boundary markers: The skill relies on structured input but does not explicitly define LLM-specific delimiters for untrusted data.
Capability inventory: Includes dxs configuration upsert for persistent configuration changes and dxs odata execute for performing data queries.
Sanitization: The documentation (specifically in references/odata-datasources.md) identifies injection risks and mandates the use of $utils.odata formatters (e.g., formatNumber, formatString) to properly escape and validate all interpolated parameters, which is an industry-standard security practice.

datasource-creator