story-context

To maintain context consistency across multi-task workflows, agents read a shared story-context file before starting a task and append factual updates after completion. This shared record prevents naming drift and inconsistent architectural decisions across tasks sharing an epic.

The best way to prevent naming drift in cross-task storytelling is enforcing a shared context file with explicit sections for Types Created, Naming Patterns, and Architecture Decisions. Agents append to this record after each task to maintain continuity.

Structure the shared context file as a short, factual markdown document with explicit sections for Types Created, Naming Patterns, and Architecture Decisions. The file must be located at planning/notes/[epic-name]/story-[story-id]-context.md and kept under 80 lines.

Cross-task context gets lost because tasks claim, implement, and move on without retaining prior decisions. A lightweight context file pattern solves this by providing a shared record that agents read before starting and append to after completion, ensuring continuity.

Multi-task workflow automation can operate without a shared context file, but it risks inconsistent decisions and naming drift. Adopting a read-before-start and append-after-completion file pattern ensures consistent patterns and architectural decisions across tasks sharing an epic.

You need a context file for multi-task storytelling workflows when multiple tasks share an epic and require consistent naming, patterns, and architectural decisions. It provides a shared record that agents read before starting and append to after completion to prevent context loss.