bonfire

To coordinate AI agents for parallel software development, you can use an orchestration skill that generates an orchestrator prompt from a plan file and manages agent dispatch, review, and remediation workflows automatically.

Agent orchestration handles crash recovery by creating plan directories with persistent state and event log files, allowing the development workflow to resume seamlessly after an unexpected interruption or crash.

To start agentic development using a high-level project plan, provide your plan file as input to the orchestration skill, which will then generate the necessary coordinator prompts and manage the execution loop.

The best way to automate a development team workflow using AI agents is to use a plan-driven orchestration approach that assumes the coordinator role and manages parallel agent execution from initial setup to task completion.

You need to provide a plan file, such as a markdown file, which the orchestrator reads to generate execution prompts and manage the state and event logs required for coordinating the development agents.

If a dispatched agent requires remediation during software development, the orchestrator manages the review and remediation workflows, coordinating corrections to ensure the final project meets the plan's objectives.