Forge app implementation patterns

This post introduces several implementation patterns for Atlassian Forge apps: 1. Code sharing: A shell script-based approach to synchronize code (utilities, types) between different code areas (Custom UI, backend functions). 2. Configuration distribution: Loading configuration data upfront in each activity, making API calls to gather Jira entity IDs, and using this configuration object for subsequent API calls. Includes a tamper check hash for security. 3. AI form helper: Using the Forge LLM API to pre-fill lengthy forms, with user review and editing capabilities. 4. Long running UI operations: Performing backend tasks in chunks, passing state between asynchronous tasks, and storing state in work items for progress reporting. Results are pushed to the UI via Forge's realtime API. 5. Resolver function keys: Using TypeScript types to define and reference resolver identifiers, ensuring singular maintenance and type safety.