The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the vendor-provided dxs CLI tool to perform lifecycle operations such as generating, validating, and uploading function configurations.
[EXTERNAL_DOWNLOADS]: The skill fetches existing function source code and configuration metadata from the Datex Studio platform using the dxs function get command.
[PROMPT_INJECTION]: The skill ingests untrusted data from existing function code and external requirements briefs produced by the requirements-gathering skill. This data is processed by the agent to generate new TypeScript code and CLI commands, representing an indirect prompt injection surface.
Ingestion points: Existing code extracted from the platform (Phase 1) and the requirements brief in the conversation context.
Boundary markers: No specific delimiters or "ignore instructions" markers are used when interpolating external content into the workflow.
Capability inventory: The skill has the ability to execute shell commands (dxs), write to the local file system (temporary .ts and .json files), and upload code to the platform (dxs function upsert).
Sanitization: No explicit validation or sanitization of the ingested code or requirements is performed before they are used to influence agent output.

function-creator