sub-agent

Multi-step code refactoring can be delegated by defining a structured task with ROLE, DIRECTIVE, CONSTRAINTS, PROCESS, and OUTPUT. This orchestrates autonomous execution and iterative refinement, reducing manual toil across multiple files.

Task delegation requires a structured format comprising ROLE, DIRECTIVE, CONSTRAINTS, PROCESS, and OUTPUT. You can also include optional resources like references and scripts to guide the autonomous workflow inside isolated environments.

Yes, you can use this sub-agent for automated code reviews and Git operations. It supports planning, execution, and iterative refinement across these tasks, enabling safe and repeatable workflows within isolated environments.

The best way to automate complex research and documentation is through structured task delegation. By providing a directive and constraints, the autonomous sub-agent iterates until the result is satisfactory, minimizing cognitive load.

Yes, this task delegation approach works for generating presentations. You define the role, directive, and output format, allowing the autonomous sub-agent to handle the multi-step planning and execution needed to create them.

You should not use an autonomous sub-agent for task delegation if your task lacks clear constraints or cannot be defined using the required ROLE, DIRECTIVE, and OUTPUT format, as structured input is essential for isolated execution.