aim-parzival-team-builder

Designing an agent team for parallel code execution requires analyzing work units for independence and mapping file ownership with zero conflicts. This skill automates that process by selecting optimal team tiers, assigning roles, and establishing Git Worktree isolation or exclusive file ownership strategies.

Structuring parallel agents for a large software project involves determining single-agent, 2-tier, or 3-tier architectures based on task complexity. This skill selects appropriate models like Sonnet for implementation and Opus for complex tasks, assigning unique cross-session memory identities to each agent.

File ownership mapping for simultaneous agents works by assigning specific files to specific roles to ensure zero conflicts. The skill assembles context blocks for each agent including ROLE, TASK, FILES, and CONSTRAINTS, applying ranked conflict avoidance strategies like Git Worktree Isolation.

Yes, you can use predefined developer roles such as analyst, architect, dev, and UX designer in an automated team workflow. The skill assigns these specific roles from a predefined set, sizes the teams accordingly, and defines agent identities with unique AI_MEMORY_AGENT_ID values.

You need a 3-tier agent team instead of a single agent when work requires complex coordination across multiple independent modules. The skill analyzes the work to be parallelized, identifies independent units, and determines the optimal tier selection based on task dependencies and file ownership requirements.

Limitations when parallelizing work across Claude Code teams include potential file ownership conflicts and dependency bottlenecks between independent units. The skill mitigates these by performing a pre-delivery review and checking for file ownership conflicts before assembling the final agent context blocks.