temporal

This skill acts as a specialized assistant for migrating legacy Netflix Conductor JSON-based workflows into modern, type-safe Temporal Python projects. It is designed for engineers transitioning from declarative Conductor DSLs to code-first orchestrations using the Temporal Python SDK. The agent automates the end-to-end lifecycle—from initial analysis of Conductor task primitives to generating robust, production-ready project structures, including activities, workflow definitions, and startup scripts.

Beyond migration, the skill provides a comprehensive suite of tools for the Temporal development lifecycle. It handles local server orchestration via temporal server start-dev, manages worker process lifecycles, and provides deep observability into workflow execution. Users can monitor worker health, inspect workflow task histories, and troubleshoot non-determinism errors or stalled activities directly from the development environment.

• Automated migration pipeline: Parses Conductor JSON definitions to generate idiomatic Temporal Python code, including type-safe activities and workflow logic.

• Local server management: Simplifies development with scripts to ensure Temporal server connectivity and clean workspace resets.

• Worker lifecycle control: Smart management of background worker processes, ensuring stale workers are killed and fresh instances are deployed during validation cycles.

• Observability and Troubleshooting: Built-in tools for listing workers, tailing logs, checking workflow status, and parsing error histories to identify stalled tasks vs. activity failures.

• Project scaffolding: Generates complete project structures with uv configuration, CI/CD-ready documentation, and standardized setup scripts for consistent deployment.

• Target Audience: Backend and Platform Engineers migrating complex stateful microservices or workflow-driven applications.

• Operational Constraints: Requires Python 3.11+ and the uv package manager; designed primarily for local development (gRPC port 7233) and individual migration tasks.

• Workflow Monitoring: Essential for identifying the difference between workflow task stalls and activity retries; always capture the Workflow ID from the starter output for precise tracking.

• Cleanup: Uses dedicated scripts (kill-worker.sh, kill-all-workers.sh) to prevent resource leakage when managing multiple local worker processes.

• Integration: Primarily acts as a plugin within the Claude Code environment, facilitating iterative refactoring and autonomous bug fixing during the migration process.