blueprint

To turn a high-level objective into a multi-session plan, provide your goal to generate a structured, step-by-step construction plan. This plan ensures memory persists across multiple sessions and coordinated sub-agents.

Managing dependencies and parallel execution across coding agents is handled using a built-in dependency graph and parallel step detection. These features optimize construction plans for coordinated work across multiple agents.

You can modify a multi-agent project plan by splitting, inserting, skipping, or reordering steps using a plan mutation protocol. This ensures the construction plan adapts safely to changes during execution.

This project-management approach requires no executable code or dependencies to run. It is a zero runtime risk, Markdown-based skill that generates construction plans without executing any code.

The best way to ensure safety and correctness in complex multi-PR migrations is using an adversarial review gate within a five-phase pipeline. This process validates plan quality before execution across multiple sessions.

You need a dependency graph for multi-agent automation tasks when executing complex migrations or coordinated work across sub-agents. It optimizes parallel execution and ensures memory persists between sessions.