implement-task

To execute tasks in parallel, this Skill spawns a team of up to four agents that work concurrently on items listed in your todos.md file, tracking progress and logging outcomes to log.md.

Yes, resumption is supported by resetting any in_progress items back to pending while preserving all completed items, allowing the agent team to seamlessly continue executing the remaining todos.md tasks.

Parallel task orchestration manages state transitions across pending, in_progress, and completed stages while automatically logging execution outcomes and progress updates to log.md within a git worktree context.

The maximum parallelism limit for spawning an agent team to execute todos.md items is four concurrent agents, with optional gating available between parallel groups to control workflow progression.

A git worktree context is required for orchestrating parallel agents, as the task execution mechanism operates within this environment to safely manage state transitions, handle failures, and monitor progress.

The task execution workflow handles failures by monitoring agent progress during parallel processing, enforcing prerequisites, and managing state transitions to ensure failed items can be resumed or re-attempted from their pending state.