AIWander

Getting started with Manager — the 48-tool multi-AI orchestrator. Use when: delegating code tasks to Claude Code/Codex/Gemini, running parallel AI executions, creating workflow templates, or routing tasks across AI backends.

Multi-vendor AI orchestration MCP server — routes coding and reasoning tasks to Claude Code, Codex, Gemini, or GPT from inside any MCP client

Wrapping manager delegations in local breadcrumbs for cross-context operation tracking when both servers are installed

Getting started with Workflow — the 37-tool automation server for API discovery, credential vault, flow recording/replay, watches, data piping, and authorized MFA helpers. Use when: automating API calls, managing credentials, recording replayable flows, polling for changes, or piping data between steps.

API pattern storage and replay, DPAPI-encrypted credential vault, flow recording/replay, scheduled watches, data transform pipelines, and trigger→action workflow chains — all through one MCP server. Teaches the graduation pipeline (browser→API), credential-by-reference pattern, flow recording, watch polling, and workflow composition. For workflow v1.1.1+.

Getting started with Hands -- the 117-tool automation server for browser, Windows desktop, and vision/OCR tasks. Use when: first time using hands, unsure which hands tool to pick, need a workflow example, or want to understand what hands can do vs other approaches.

Browser automation, Windows desktop automation, vision/OCR, and workflow graduation pipeline -- all through one MCP server. Teaches the escalation ladder, a11y_ref pattern, batch operations, API discovery pipeline, UIA desktop control, and vision-as-verification. For hands v1.1.1+.

Windows-native MCP server — shell, files, sessions, transforms, and breadcrumb operation tracking

Getting started with local — the 105-tool Windows filesystem, shell, and session server. Use when: first time using local, need file I/O, shell commands, git ops, transforms, PowerShell sessions, SQLite queries, or Windows registry access.

Canonical skill for the echo MCP server (v1.1.1). Local-first, model-agnostic AI toolkit wrapping Ollama: embeddings, semantic search, generation, summarization, pattern matching, and HTML fetch+summarize. Every tool requires the model as a per-call parameter. No defaults, no hardcoded model names, no assumptions.