Install, configure, use, debug, and troubleshoot kagent OSS and Solo Enterprise for kagent on Kubernetes. Use when Codex needs to author or review kagent manifests, Helm values, model and MCP server configuration, agent prompts or skills, or diagnose runtime, authn, and authz issues across OSS and Enterprise deployments, including AccessPolicy, OIDC, management/workload topology, and repo-versus-doc API drift.
SDLC workflow with MCP tools. Triggers on "start", "implement", "work on", or unclear workflow.
Helps with Gate Pay x402 payments (HTTP 402), finding payable merchants when discovery is available, and wallet setup: Quick Wallet, Gate plugin token, local EVM key, or Gate Exchange. Use when the user pays or wants to pay with x402 or 402, configures a Gate or MCP wallet or payment default, lists merchants or paid services, orders something without a merchant link, or says 快捷钱包, 插件钱包, 私钥, 支付方式, 有哪些商户, 能付费的服务列表, 配置钱包, gate钱包, MCP钱包, 交易所, 下单, 支付, or PAYMENT_METHOD_PRIORITY. Do NOT use for trading-only or market chat with no Gate Pay x402 or wallet intent, or for completing the same Gate Pay order on another vendor's x402 MCP.
Analyze PyTorch internals across Python, C++, and CUDA layers using the TorchTalk MCP server. Use when asked about how PyTorch operators work internally, where functions are implemented, what would break if code is modified, or finding tests for PyTorch operators.
Author Grafana dashboards with MCP-driven metric discovery, visual iteration via the image renderer, and consistent layout conventions extracted from existing dashboards. Use when: (1) Creating new Grafana dashboards, (2) Modifying existing dashboard JSON, (3) Adding panels or sections to dashboards, (4) Choosing metrics and PromQL for panels, (5) Debugging blank or broken dashboard panels, (6) Laying out dashboard grids. Triggers: "grafana dashboard", "create dashboard", "add panel", "dashboard layout", "grafana json", "dashboard ConfigMap", "new dashboard", "visualize metrics"
Manage Model Context Protocol (MCP) servers - discover, analyze, and execute tools/prompts/resources from configured MCP servers. Use when working with MCP integrations, need to discover available MCP capabilities, filter MCP tools for specific tasks, execute MCP tools programmatically, access MCP prompts/resources, or implement MCP client functionality. Supports intelligent tool selection, multi-server management, and context-efficient capability discovery.
Master of text-based visual communication using ASCII art, Unicode box-drawing, and structured diagram notation. Creates clear, maintainable diagrams for systems, processes, hierarchies, relationships, and psychological structures. Proactively generates diagrams to enhance understanding. Activate on visualization needs, system architecture, process flows, psychological mapping, or when complex concepts would benefit from visual representation. NOT for photo editing, vector graphics, or GUI-based design tools.
OpenAI Responses API for stateful agentic applications with reasoning preservation. Use for MCP integration, built-in tools, background processing, or migrating from Chat Completions.
Anthropic Claude Agent SDK for autonomous agents and multi-step workflows. Use for subagents, tool orchestration, MCP servers, or encountering CLI not found, context length exceeded errors.
OpenAI Agents SDK for JavaScript/TypeScript (text + voice agents). Use for multi-agent workflows, tools, guardrails, or encountering Zod errors, MCP failures, infinite loops, tool call issues.
Build AI agents on Cloudflare Workers with MCP integration, tool use, and LLM providers.
Activate when users ask about Claude Code installation, slash commands (/cook, /plan, /fix, /test, /docs, /design, /git), creating/managing Agent Skills, configuring MCP servers, setting up hooks/plugins, IDE integration (VS Code, JetBrains), CI/CD workflows, enterprise deployment (SSO, RBAC, sandboxing), troubleshooting authentication/performance issues, or advanced features (extended thinking, caching, checkpointing).
Search the web and code context via Exa. Use when you need Exa web search, Exa code context, or Exa MCP tools. Triggers on "Exa", "web search", "search the web", "code context", "mcp exa".
Search real-world code examples from millions of GitHub repositories using literal patterns. Use when users need to find implementation examples, API usage patterns, or production code snippets. Triggers on "search code for", "find code examples", "how do developers use", "show me real examples of", "grep code".
Control JetBrains IDE programmatically for file operations, code analysis, refactoring, and terminal execution. Use when users want to interact with IntelliJ/WebStorm/PyCharm, run configurations, refactor code, or analyze files. Triggers on "open file in IDE", "run configuration", "refactor", "find files in project", "IDE terminal".
Automate Chrome DevTools browser actions like navigation, forms, screenshots, and performance traces. Use when users need browser automation or page inspection. Triggers on "Chrome DevTools", "take a screenshot", "fill out the form", "performance trace", "click on".
Manage MCP servers - discover, analyze, execute tools/prompts/resources. Use for MCP integrations, capability discovery, tool filtering, programmatic execution, or encountering context bloat, server configuration, tool execution errors.
Dynamic MCP server discovery and code-mode execution via central registry. Use for multiple MCP integrations, tool discovery, progressive disclosure, or encountering MCP context bloat, changing server sets, large tool sets.
