3.6 Run the First Complete Job

A complete task can start with inputs already prepared in BioLine, run a suitable upstream Flow, and then pass successful outputs to a downstream BioLine Tool.

Example request to OpenClaw:

“I uploaded paired-end FASTQ, a host reference genome, and metadata.tsv. Discover a suitable BioLine Flow, inspect its input ports and parameters, and show me the plan. Wait for my confirmation before submitting. After upstream success, discover a suitable downstream Tool for diversity analysis, differential analysis, and reporting, and record the exact version and resource requirements.”


Typical stages are:

• Check the BioLine data directory, sample naming, and metadata.
• Discover and inspect candidate Flows with fkit, then bind inputs from the actual response.
• Present the Flow and Version IDs, parameters, defaults, bindings, and impact. Create the Job only after confirmation.
• Poll the BioLine Job; after success, download and finalize outputs into stage.
• Discover a suitable downstream BioLine Tool and inspect the exact version’s summary/config, downstream marker, and resources.
• After confirmation, pull that version on demand, generate and save a task script, and run it in an isolated environment with networking disabled.
• Retain the Tool version, documentation snapshot or hash, pull identity, script, command, logs, and results with the task.

Reuse depends on live discovery. If no suitable Tool or Flow exists, use the separate authoring and confirmation workflow instead of guessing an ID or silently switching versions.