cy-workflow-memory

To maintain workflow memory across multiple AI tasks, load shared and task-local memory files before code generation, then apply compaction rules and promotion decisions to keep cross-task context durable and synchronized.

Memory compaction in automated code generation workflows is the process of applying strict rules to consolidate and reduce task-local memory files, preventing context loss and noisy notes while preserving durable cross-task information.

To update task memory when objectives shift during code generation, dynamically write to task-local memory files whenever decisions are made or errors occur, ensuring shared workflow memory stays synchronized with the current state.

Yes, the workflow memory skill requires specific directory paths to operate, validating provided memory directories and files before execution to resolve repository conflicts and safeguard against accidental data loss.

The best way to prevent context loss in multi-step AI code tasks is to enforce promotion tests and compaction rules on workflow memory files, ensuring durable cross-task context remains reliable and up-to-date throughout execution.

AI task memory gets duplicated or filled with noisy notes when compaction rules and promotion decisions are not enforced, causing unmanaged updates to overwrite or clutter shared workflow memory during multi-step execution.