labtasker

What problem does it solve?

This skill addresses the lack of built-in Labtasker knowledge in AI coding agents, removing the need for agents to manually reference external documentation to help users write valid task submission, scheduling, and management scripts for machine learning experiments.

Core Features & Use Cases

• Dual API Script Generation: Teaches agents to create both Bash CLI and Python API scripts for Labtasker task operations, matching user preferred workflows.
• End-to-End Task Management: Enables agents to implement workflows for submitting, filtering, updating, deleting, and monitoring experiment tasks, including handling task failures and resuming interrupted runs.
• Event-Driven Automation: Guides agents to write scripts for real-time event listening, such as sending email notifications when tasks fail.
• Use Case: An ML engineer can ask their AI agent to write a script that submits 500 image training tasks with custom metadata, filters for tasks that failed due to out-of-memory errors, and automatically resubmits them with adjusted batch sizes, no external documentation lookup required.

Quick Start

Instruct your AI coding agent to generate a Labtasker Python script that submits experiment tasks with custom arguments and metadata, then filters for failed tasks to resubmit with modified parameters.