Skills, MCP and AI agents

Package reusable instructions as Skills, connect tools through MCP and delegate bounded tasks with clear inputs, permissions and review criteria.

Build agent workflows around a concrete task. Choose reusable Skills, expose the tools it needs through MCP and define what each agent must deliver before integration.

Markdown

• AI first company meaning: how agents change business processesAI first company examples: marketing, support, analytics and coding tasks, plus six steps for a transition with human oversight.
• Google Antigravity Remote Control: How to Control the Agent from Your PhoneHow to enable Remote Control in Google Antigravity on desktop and in the CLI: the toggle, the agy remote-control daemon, and what to do if you have connection problems.
• Dots in ChatGPT: What They Are, How to Create Your Own Agent, and How to Use ItHow to create a dot in ChatGPT, give it its first task, configure access, check the result, and stop tasks and schedules separately.
• What Is an AI Agent in Plain English, and Where Can You Use One?AI agents in plain English: how they differ from chatbots, what they consist of, what they can do on their own, and when to use one—and when not to. With examples.
• Claude Code Is Slowing Down: How to Choose Skills, Subagents, MCP, and PluginsA comparison of Claude Code extensions, context usage, seven mistakes, and a minimal working setup, with technical examples preserved.
• Claude API Agent Skills: connect a ready-made skill and download its fileConnect a pptx skill with container.skills and code execution, inspect the response, and download its presentation through the Claude Files API.
• Git worktree in Claude Code: parallel sessionsHow to use --worktree, .worktreeinclude, worktree.baseRef, and subagent isolation. Differences from branches, cleaning up worktrees, and common mistakes.
• Claude Code and Codex Together: 3 Work PatternsWorking together with Claude Code and Codex: writing and reviewing code, drafting and architecture, separating infrastructure and business logic, synchronizing instructions.
• Claude Code: How to Reconfigure an Installed Plugin Without Reinstalling It (2.1.285)Claude Code 2.1.285 introduced the claude plugin configure command. How to check plugin settings in the documentation and change a token or address without reinstalling.
• Claude Code Projects: coordinator, threads, and beta limitationsHow Claude Code Projects cloud threads work: GitHub App, instructions, memory, conflicts, and beta limitations.
• Claude Code Remote Control: How to Code from Your Phone with a QR Code in 2026Step-by-step instructions for connecting Claude Code to your phone using /rc and a QR code. Commands, limits, security, and 6 common first-run bugs.
• How to Control Claude Code from Your Phone via Telegram in 2026Step by step: create a Telegram bot, install the Claude Code Channels plugin, secure it with an allowlist, and run it 24/7 on a VPS. A ready-to-follow guide with commands.
• Claude Code Tool Search: How the AI Agent Finds an MCP ToolHow tool search works in Claude Code: what is loaded into context immediately, when the mechanism turns itself off, and how to configure ENABLE_TOOL_SEARCH and alwaysLoad.
• What Is Cowork in Claude, and How Is It Different from Claude Code?Chat, Cowork, and Claude Code: working with folders, business tasks, confirmations, and technical limitations.
• How to Connect Codex, Claude Agent, and OpenCode to Devin Desktop via ACP in 2026How to connect Codex, Claude Agent, and OpenCode in the Devin Desktop editor through the ACP protocol. Installation commands, settings, and plans.
• Codex on your phone: how to continue a task through Codex CloudCreating and publishing a Codex cloud environment, continuing a task from your phone, updating the environment, resources, and limitations.
• Moving a project from Codex to Hermes on a serverTransferring a repository to a server to work in Hermes: saved changes, dependencies, secrets, the working folder, and checks before continuing development.
• Codex Skills: Creating a Skill for Repetitive TasksSKILL.md structure, loading paths, explicit and automatic invocation, result criteria, and transferring a skill to Claude Code.
• Codex worktree: a separate working copy for a second taskWorking copy of Codex CLI 0.155.1: experimental worktree flag, source HEAD, checking the new folder, preserving the result, and safe removal.
• Cursor Projects: what it is and how the coordinator delegates tasksCursor Projects, as announced by Cursor: how the coordinator delegates tasks, where shared context is stored, how Subscriptions work and what this beta is for.
• How to Launch Hundreds of Agents in Claude Code with Dynamic WorkflowsAnthropic launched Dynamic Workflows on 28.05.2026. How Claude writes scripts for 1,000 subagents, what ultracode and /deep-research are, and the limits.
• /goal in Claude Code: Autonomous Mode and Agent View in 2026The /goal command in Claude Code 2.1.139 launches autonomous mode: the agent works until it achieves the goal. I explain Agent View, worktrees, and 5 examples.
• How to Run Claude Code and Codex Overnight via /goal in 2026: 14-Hour Autonomous ModeThe /goal command in Claude Code 2.1.139 and Codex CLI runs an AI agent autonomously for hours. I explain the setup, real examples, and how many tokens it costs.
• Grok Build memory across sessions: settings and limitationsHow to enable memory, where notes are stored, and how /flush, /remember, /dream, and /memory work. Settings priority, checking, and deleting entries.
• AI Agent Hermes: What It Is, How to Install and Configure ItHermes Agent in plain language: what this AI agent is, how much it costs, how to install it on Windows, macOS, and Linux, choose a model, and connect Telegram.
• Hermes Agent with a local model: connecting OllamaHow to connect a local model to Hermes Agent: Local Models in the desktop app or Ollama, a model with tool calling, 64,000 context, and how to check that it works.
• How to keep Hermes Agent running on a server around the clockHow to keep Hermes Agent running on a server 24/7: installation on a VPS without a browser, running the gateway as a system service, starting it automatically after a reboot, linger, and Docker.
• AI Agent Skills in Hermes: Install a Skill and Create Your OwnDirectories and GitHub, the SKILL.md format, /learn, checking, updating, and removing Hermes Agent skills with complete command examples.
• How to Connect Hermes Agent to Telegram and Manage Tasks from Your PhoneStep-by-step setup: create a bot through BotFather, add the token and access list in Hermes Agent, and start the gateway. Voice messages, scheduled tasks, and phone control.
• Hermes profiles: separate personal and work agentsHow to create personal and work profiles in Hermes Agent without cloning, configure each separately, and check their paths. Profile boundaries: memory, sessions, files, and accounts.
• Create Images and Videos with an AI Agent: Higgsfield MCP in Claude CodeHow to connect Higgsfield MCP to Claude Code with one command, give the agent a task to create an image or video, and when to turn off the server to avoid using up limits.
• Claude Code Hooks Documentation: Setup and 7 Ready-to-Use ExamplesClaude Code hooks docs in practice: four handler types, lifecycle events, settings.json, seven examples, testing, and common pitfalls.
• Plugins for an AI agent Telegram bot: installation and checkingWhere a plugin is installed, how it differs from a skill and MCP, how to reload the agent, and how to check commands from Telegram.
• How to Choose Claude Code Skills in 2026: Context and SelectionHow to choose Claude Code skills: context budget, usage frequency, security, overlap, measurability, and a short list of vetted skills.
• How to Configure Claude Code So It Doesn’t Break Your Project Overnight7 hooks, settings.json with permissions, and a 12-point checklist to leave Claude Code working overnight without losing .env or a database. Protection in 4 layers.
• AI Agents for Business in 2026: Which Tasks Can You Really Hand Off?Three task buckets: what an AI agent for business handles end to end, what it can do under supervision, and what it consistently fails at. Anthropic's figures and a one-week test.
• One-Person Business: How to Replace Staff with AI in 2026How to divide roles between a person and AI agents, start with one process, retain accountability, and assess the costs of a one-person business.
• Claude Code overnight: loop engineering and Ralph Loop in 2026How to run Claude Code in an autonomous loop: Ralph Loop, /loop, /goal, /batch. Ready-to-use scenarios and protection against burning through tokens in 2026.
• Claude Code MCP configuration file: project scope and 7 serversClaude Code MCP config file location: choose project scope for .mcp.json, connect servers with claude mcp add, and verify them in /mcp.
• Parallel Claude Code Work: worktree, Agents, and WorkflowsFour layers of parallel Claude Code work, git worktree isolation, outcome verification, GUI tools, and limits on workload and tokens.
• Multi-Agent Systems: What They Are, in Plain Language, in 2026Multi-agent systems in plain language: how several AI agents differ from one, how an orchestrator and workers operate, and when the combination pays off.
• n8n vs Claude Code in 2026: what to choose for automationA comparison of n8n and Claude Code: when no-code hits its ceiling, how to migrate a working flow to an AI agent, pricing, and 6 signs it’s time to move on.
• OpenClaw: what it is, how to install it, and connect it to TelegramOpenClaw in plain language: what this personal AI assistant is, how to install it on Windows, macOS, and Linux, choose a model, and connect Telegram.
• Routines in Claude Code: 5 ready-to-copy routinesRoutines in Claude Code 2026: 3 trigger types (cron/webhook/GitHub), 5 ready-to-copy routines, plan limits, and step-by-step setup.
• Claude Code Skills in 2026: how to build a skills library for yourselfHow to create Claude Code Skills: SKILL.md structure, the description formula, ready-made collections from Anthropic, and anti-patterns. Step by step in 30 minutes.
• AI Agent Skills, Subagents, MCP, and Plugins in Claude Code: Which to ChooseSkills, subagents, MCP, and plugins are four layers of Claude Code extensions. I explain when to use each, with a ready-made decision matrix and project examples.
• Claude Code Subagents: Structure, Five Roles, and OrchestrationAgent file, description, five complete templates, parallel execution, Managed Agents, limitations and cost with examples preserved.
• Ultracode in Claude Code 2026: how to enable it with one commandUltracode in Claude Code: a mode where Claude launches a team of agents itself. How it differs from /effort xhigh, how to enable it, and how quickly it burns through tokens.
• How to delegate a large task in Claude Code using nested subagentsIn v2.1.172, Anthropic enabled Claude Code subagents to spawn their own subagents up to 5 levels deep. Instructions on delegating a large task.
• A Second Brain in Claude Code: How to Build the Structure in an HourHow to build a second brain for Claude Code: a memory hierarchy, a five-part folder structure, and git so the agent remembers everything about your project between sessions.
• WhatsApp Business Tools MCP: how to connect WhatsApp Business to an AI agentMeta's official MCP server for WhatsApp Business: what the agent can do, the connection command in Claude Code, setup in Codex and ChatGPT, and beta version limitations.
• How to build a live business dashboard in ClaudeConnectors, Cowork, and published artifacts: a ready-to-use dashboard prompt, metrics, limitations, and read-only access.