AgentToolHub
Skill

cascade-workflow

cascade-workflow — a Claude Code skill from rysweet/amplihack

Features

  • ✓**[dev-orchestrator (`/dev`)](/dev)** — Unified task orchestrator with
  • ✓**[Recipe Runner](docs/recipes/README.md)** — Code-enforced workflows (10
  • ✓**[Auto Mode](https://rysweet.github.io/amplihack/AUTO_MODE/)** — Autonomous
  • ✓**[Multitask](~/.amplihack/.claude/skills/multitask/SKILL.md)** — Parallel
  • ✓**[Expert Panel](/amplihack:expert-panel)** — Multi-expert review with voting
  • ✓**[N-Version Programming](/amplihack:n-version)** — Generate multiple
  • ✓**[Document-Driven Development](https://rysweet.github.io/amplihack/document_driven_development/)**
  • ✓**[Pre-Commit Diagnostics](~/.amplihack/.claude/agents/amplihack/specialized/pre-commit-diagnostic.md)**
  • ✓**[CI Diagnostics](~/.amplihack/.claude/agents/amplihack/specialized/ci-diagnostic-workflow.md)**
  • ✓**[Cascade Fallback](/amplihack:cascade)** — Graceful degradation
  • ✓**[Quality Audit](/amplihack:analyze)** — Seek/validate/fix/recurse quality
  • ✓**[Kuzu Memory System](https://rysweet.github.io/amplihack/AGENT_MEMORY_QUICKSTART/)**
  • ✓**[Investigation Workflow](#workflows)** — Deep knowledge excavation with
  • ✓**[Discoveries](https://rysweet.github.io/amplihack/DISCOVERIES/)** —
  • ✓**[Knowledge Builder](/amplihack:knowledge-builder)** — Build knowledge base
  • ✓**[Goal-Seeking Agent Generator](https://rysweet.github.io/amplihack/GOAL_AGENT_GENERATOR_GUIDE/)**
  • ✓**[GitHub Copilot CLI](https://rysweet.github.io/amplihack/COPILOT_CLI/)** —
  • ✓**[Microsoft Amplifier](https://github.com/microsoft/amplifier)** —
  • ✓**[RustyClawd](#rustyclawd-integration)** — High-performance Rust launcher
  • ✓**[Remote Execution](~/.amplihack/.claude/tools/amplihack/remote/README.md)**
  • ✓**[Security Analysis](/amplihack:xpia)** — Cross-prompt injection defense
  • ✓**[Socratic Questioning](/amplihack:socratic)** — Challenge claims and clarify
  • ✓**[Benchmarking](https://rysweet.github.io/amplihack/BENCHMARKING/)** —
  • ✓**[Customization](/amplihack:customize)** — User preferences (verbosity,
  • ✓**[Statusline](https://rysweet.github.io/amplihack/reference/STATUSLINE/)** —
  • ✓**Scout** discovers all VMs and sessions via azlin (no SSH needed for
  • ✓**Admiral reasoning** uses LLM streaming to decide: wait, send_input, restart,
  • ✓**SessionCopilot** watches local sessions and auto-continues toward a goal
  • ✓**Dual backend** — uses Anthropic API when available, falls back to GitHub
  • ✓**Safety** — dangerous input patterns blocked, shell metacharacter rejection,

Installation Guide

Install the prerequisites above first, then choose an option below.

Option 1: Quick Start (requires uv and Python)

bash
# Launch with Claude Code
uvx --from git+https://github.com/rysweet/amplihack amplihack claude

# Launch with Microsoft Amplifier
uvx --from git+https://github.com/rysweet/amplihack amplihack amplifier

# Launch with GitHub Copilot
uvx --from git+https://github.com/rysweet/amplihack amplihack copilot

This launches an interactive Claude Code session enhanced with amplihack's
workflows, specialized agents, and development tools. You'll get a CLI prompt
where you can describe tasks and the framework orchestrates their execution.

Rust CLI status

The main amplihack entrypoint is still Python-owned today, but the runtime is
no longer purely Python-first:

  • when an installed Rust CLI is available, the Python entrypoint now hands

    • install, recipe, mode, and update through to that Rust binary
  • when amplihack-hooks is installed, hook configuration now prefers the Rust

    • hook engine automatically unless AMPLIHACK_HOOK_ENGINE=python is set

    The fully isolated helper below is still the easiest way to exercise the Rust
    CLI end-to-end without touching your real ~/.claude state. On a fresh machine,
    it downloads the latest compatible published amplihack-rs release binary
    automatically.

    bash
    # Single-command GitHub install for daily use
uvx --from git+https://github.com/rysweet/amplihack \
  amplihack-rust-trial \
  install

# The installer downloads the latest published Rust release, places it in
# ~/.local/bin, updates your shell profile PATH, and runs `amplihack install`
# automatically so the framework assets are ready immediately.

# Fully isolated fresh-machine flow
uvx --from git+https://github.com/rysweet/amplihack \
  amplihack-rust-trial \
  --trial-home ~/.amplihack-rust-e2e \
  copilot

# Try specific commands without touching your current amplihack install
uvx --from git+https://github.com/rysweet/amplihack amplihack-rust-trial recipe list
uvx --from git+https://github.com/rysweet/amplihack amplihack-rust-trial mode detect

    By default the helper stores trial state under ~/.amplihack-rust-trial. It
    remains useful for proving end-to-end Rust behavior because the default product
    entrypoint and install/bootstrap flow are still not fully Rust-native.

    Option 2: Global Install (for daily use)

    bash
    # Install once directly from GitHub
uvx --from git+https://github.com/rysweet/amplihack amplihack-rust-trial install

# Use directly
amplihack claude
amplihack amplifier
amplihack copilot

# Update later
amplihack update

    First Session

    After launching amplihack (e.g., amplihack claude), you'll be inside an
    interactive agent session — a chat interface powered by your chosen coding
    agent. Everything you type in this session is interpreted by amplihack's
    workflow engine, not by your regular shell.

    New users — start with the interactive tutorial:

    text
    I am new to amplihack. Teach me the basics.

    This triggers a guided tutorial (60-90 minutes) that walks you through
    amplihack's core concepts and workflows.

    Experienced users — just describe what you want to build:

    text
    cd /path/to/my/project
Add user authentication with OAuth2 support

    The /dev command is amplihack's primary entry point for development tasks. It
    automatically classifies your task, detects parallel workstreams, and
    orchestrates execution through the 23-step default workflow.

    Developer Quick Example

    Here is a complete end-to-end example of amplihack in action:

    1. Single task — fix a bug:

    bash
    cd /path/to/your/project
/dev fix the authentication bug where JWT tokens expire too early

    What happens:

  • Classifies as: Development | 1 workstream

  • Builder agent follows the full 23-step DEFAULT_WORKFLOW

  • Creates a branch, implements the fix, creates a PR

  • Reviewer evaluates the result — if incomplete, automatically runs another

    • round
  • Final output: # Dev Orchestrator -- Execution Complete with PR link

    • 2. Parallel task — two independent features at once:

    bash
    /dev build a REST API and a React webui for user management

    What happens:

  • Classifies as: Development | 2 workstreams

  • Both workstreams launch in parallel (separate /tmp clones)

  • Each follows the full workflow independently

  • Both PRs created simultaneously

    • 3. Investigation — understand existing code before changing it:

    bash
    /dev investigate how the caching layer works, then add Redis support

    What happens:

  • Detects two workstreams: investigate + implement

  • Investigation phase runs first, findings pass to implementation

  • Result: informed implementation with full context

    • What you'll see during execution:

  • [dev-orchestrator] Classified as: Development | Workstreams: 2 — starting execution...

  • Builder agent output streaming (the actual work)

  • Reviewer evaluation with GOAL_STATUS: ACHIEVED or PARTIAL

  • If partial — another round runs automatically (up to 3 total)

  • # Dev Orchestrator -- Execution Complete with summary and PR links

    • Note: The Task() syntax shown in some documentation is an advanced

    programmatic API for scripting agent workflows. For interactive use, plain

    natural language prompts are all you need.

    Reviews