swarm-planner

Multi-agent task planning prevents conflicts by generating a structured task graph with explicit dependencies. Tasks are defined atomically and reviewed by subagents, ensuring coordinated parallel execution across software projects without overlapping work.

You generate a task dependency graph by providing project goals and a directory. The Skill fetches current library documentation, enumerates atomic tasks, maps dependencies, and outputs a structured plan with validation criteria for architecture upgrades.

Yes, the Skill uses Context7 to fetch current documentation for libraries and frameworks before planning. This ensures the generated dependency-aware implementation plan is based on accurate, up-to-date technical context.

The best way to ensure atomic, reviewable tasks is to enforce guardrails and a pre-yield review. The Skill requires every task to be atomic and undergoes a review to ensure coverage and robustness before yielding the structured plan.

Yes, plan-driven orchestration works both inside and outside Plan Mode. The Skill generates explicit-task dependency plans and outputs a structured format with validation criteria regardless of the mode you choose to run.

You need explicit dependency planning when coordinating parallel multi-agent execution to prevent conflicts. It ensures tasks declare dependencies, remain atomic, and are validated by subagents for robust cross-agent collaboration.