Skill UI

This guide details advanced architectural patterns for building complex AI systems using multiple Language Model instances (agents). It addresses the single-agent context bottleneck by distributing work, enabling parallelism, and facilitating specialization across different subtasks. Key patterns include Supervisor/Orchestrator, Peer-to-Peer/Swarm, and Hierarchical designs, providing guidelines for context isolation and effective coordination protocols.

Implements standardized function/tool calling across various large language models (LLMs) via OpenRouter. This unified approach allows developers to build sophisticated, multi-step agents and complex workflows, regardless of the underlying provider (e.g., GPT-4o, Claude 3.5). It simplifies the process of achieving structured output and tool-augmented reasoning in production systems.

Enables Claude to orchestrate multi-agent systems using AI SDK v5, setting up agents, handoffs, intelligent routing, and workflows across OpenAI, Anthropic, and Google when users request complex agent coordination.

Orchestrates multiple agents for complex workflows, decomposing feature, review, refactor, test, or documentation requests into parallel subtasks tracked via cc-mirror and TodoWrite for live visibility.

This skill enables the coordination of multiple specialized AI agents within a single session. It is designed for complex, multi-domain tasks such as comprehensive code reviews, security audits, or full-stack feature implementations. It coordinates expertise across various domains (backend, frontend, testing, architecture) to provide holistic analysis and actionable recommendations.

Guides orchestrating multiple Task tool subagents with run_in_background for concurrent work, covering when to trigger parallelism, how to craft prompts, launch in one message, synchronize results, and integrate TodoWrite status updates.

A sophisticated AI tool that coordinates multiple specialized agents (e.g., Security, Architecture, Performance) to provide comprehensive, multi-perspective code assessments. It goes beyond single-point reviews by managing context, running hybrid execution workflows, resolving conflicts, and synthesizing unified, actionable reports for high-quality code assurance.

Persistent file-based planning for multi-step AI-agent tasks, using task_plan.md, findings.md, and progress.md as on-disk working memory with recovery hooks and optional session catch-up.

Automates Renderform workflows via Composio's Rube MCP, guiding agents to search current schemas, validate active connections, and execute operations (including multi-tool calls) with schema-compliant inputs and session management.

This skill provides implementation patterns for setting up and managing a multi-environment connection with Retell AI. It is essential for developers building advanced AI voice agents, integrating complex phone call automation, or utilizing robust telephony workflows. It guides users through SDK usage and handles common API error scenarios.

This skill implements a robust, multi-layered security hardening process based on the Defense-in-Depth strategy and DevSecOps principles. It coordinates specialized agents to perform comprehensive security assessments (SAST, DAST, SBOM, Threat Modeling), identify critical vulnerabilities, and execute layered remediation. The process covers application code, backend APIs, frontend components, and mobile platforms, ensuring a resilient security posture across the entire development lifecycle.

SGLang is a high-performance serving framework for LLMs and VLMs that uses RadixAttention prefix caching to deliver structured outputs (JSON/regex/grammar) and agentic workflows with function calls, providing up to 5× faster inference than vLLM across multi-GPU production deployments.