worker-integration

The worker-integration skill provides a robust framework for managing interaction between background workers and specialized AI agents within the Ruflo (formerly Claude Flow) ecosystem. It acts as a central nervous system for task dispatch, ensuring that incoming triggers—such as ultralearn, optimize, audit, or refactor—are handled by the most capable agents available. By leveraging execution history, quality scores, and latency metrics, the system dynamically selects the optimal agent to maximize task success and efficiency, effectively minimizing token waste and execution time.

• Intelligent Dispatching: Automatically routes tasks to primary, secondary, or fallback agents based on trigger type and specific pipeline phases like discovery, vectorization, or security scanning.
• Performance-Based Selection: Employs a learning loop that tracks quality scores, success rates, and average p95 latency to influence future agent selection decisions.
• Consistent Memory Patterns: Enforces structured data storage using a defined {trigger}/{topic}/{phase} naming convention, facilitating seamless retrieval and context persistence across agent sessions.
• Benchmark Compliance: Monitors agent output against predefined thresholds (e.g., p95 latency, memory usage, or scan coverage) to ensure operational standards are met.
• Feedback-Driven Improvement: Provides built-in methods to record execution outcomes, which directly feed into the agent selection model to continuously optimize the swarm's performance.

This skill is intended for software engineers, platform architects, and teams using Claude Code to manage complex, multi-agent workflows. It is particularly useful for enterprise environments where maintaining code quality, security standards, and architectural decisions requires consistent, automated oversight. To get started, users can utilize the npx agentic-flow CLI to inspect agent recommendations, view real-time integration statistics, and check benchmark compliance. Practical use cases include automated security audits during CI/CD, recursive code refactoring tasks, and continuous performance benchmarking of API modules. The system assumes a JSON-based configuration inside .claude settings, allowing teams to explicitly map triggers to specialized agent roles, ensuring that developers spend less time on manual coordination and more time on core business logic.