agent-swarm

Agent Swarm is a framework for orchestrating AI coding agents. It enables teams to automate complex coding tasks, improve code quality, and accelerate development cycles. Integrates with Claude and other AI agents to streamline workflows.

[{"step":"Define the task scope. Specify the repository/project, files/modules to focus on, and the agents to deploy (e.g., code reviewer, test generator, documentation agent).","tip":"Use specific file paths or module names to avoid ambiguity. For example, instead of 'payment code,' use 'payment_service.py and stripe_integration.py.'"},{"step":"Set clear criteria for each agent. For the code reviewer, specify review criteria (e.g., security, performance, or style). For the test generator, define which functions or classes need tests. For the documentation agent, outline what to update.","tip":"Prioritize tasks by labeling them as 'critical,' 'high,' or 'low' priority to guide the agents' focus."},{"step":"Deploy the Agent Swarm. Use the orchestration command or tool (e.g., `agent_swarm orchestrate --task=\"[TASK]\" --agents=\"[AGENT_LIST]\"`).","tip":"Start with a small subset of agents or files to validate the output before scaling up."},{"step":"Review the orchestration report. Check for issues flagged by the code reviewer, test coverage improvements, and documentation updates.","tip":"Use the report to prioritize fixes and next steps. Focus on critical issues first."},{"step":"Iterate and refine. If the output isn't satisfactory, adjust the task scope or agent criteria and redeploy the swarm.","tip":"For complex tasks, break them into smaller subtasks and orchestrate agents for each subtask separately."}]