Multi-Task Switch Manager

When Parallel Work Becomes a Cognitive Tax

Managing several projects at once creates a different problem from simple task overload: repeated context switching leaves attention residue. A deep-work session gets interrupted by messages, ad-hoc requests, or details from another project, and a packed schedule may still lack clear deliverables. This skill reframes multitasking as a scheduling and switching-cost problem. It starts by collecting projects, due dates, progress, complexity, and dependencies, then identifies task pairs with high switching costs, batches similar work, and places demanding tasks in energy peaks.

Capabilities, Steps, and Limits

Core capabilities include task triage, switching-cost review, time-block allocation, an urgent-important priority matrix, buffer planning, and a parallel-project timeline view. A typical run begins with inventory: list active tasks, note due dates and current progress, mark whether each needs deep thinking or mechanical execution, and capture dependencies. It then classifies tasks with the matrix, finds batchable items, evaluates switching costs, and proposes daily blocks based on energy patterns. Outputs should use at least 30-minute blocks, keep deep-work blocks at 90 minutes or longer, reserve daily buffer, place similar tasks adjacent, and define a concrete outcome for each block.

Limits matter: it does not connect directly to Jira, Lark/Feishu, or similar tools, so task details must be provided manually. Recommendations are general, not medical or team-level project management. If more than 10 tasks run in parallel, everything is marked P0, or due dates are missing, the first step is subtraction, clarification, or dependency review rather than scheduling a full calendar.