ship-parallel

Parallel execution of GitHub issues requires dependency analysis and worktree-based concurrency to isolate branches. This skill inspects issue dependencies, classifies safe parallel sets via SM-based logic, and coordinates TeamCreate agents to prevent avoidable merge conflicts.

Agent orchestration for mixed issue sets applies SM-based parallel classification to separate independent issues from dependent ones. It orchestrates coordinated ship workflows using worktree-based concurrency, falling back to sequential handling for issues requiring strict ordering.

You can ship multiple issues simultaneously using Claude Code by preparing manual worktrees for each issue and triggering TeamCreate. The skill monitors progress via SendMessage and requires explicit user confirmation at intervention points to ensure coordinated execution.

Team-based agent coordination for parallel development is managed by assigning each issue to an individual worktree and using TeamCreate to spawn agents. SendMessage monitoring tracks progress while preset constraints and user confirmation guards prevent runaway execution.

Manual worktree preparation is required before parallel ship workflows because each issue must be isolated in its own branch to prevent file-level merge conflicts. This setup allows TeamCreate agents to operate concurrently without overwriting shared working directories.

Sequential fallback handling for issues is necessary when dependency analysis reveals conflicting file targets or direct dependencies between issues. The skill's SM-based parallel classification detects these risks and automatically routes dependent issue sets into sequential ship workflows.