Skip to main content

Modern CLI installer for nwave with rich terminal UI

Project description

nWave

AI agents that guide you from idea to working code — with you in control at every step.

nWave runs inside Claude Code. You describe what to build. Specialized agents handle requirements, architecture, test design, and implementation. You review and approve at each stage.

Quick Start

Choose your installation method. Both work equally well.

Plugin (Beta Preview)

From Claude Code, run:

/plugin marketplace add nwave-ai/nwave
/plugin install nw@nwave-marketplace

Restart Claude Code and type /nw: to see all available commands.

CLI Installer (Recommended)

Install from PyPI with no repository clone needed:

pipx install nwave-ai
nwave-ai install

Agents and commands go to ~/.claude/.

Don't have pipx? Install with: pip install pipx && pipx ensurepath, then restart your terminal. Windows users: Use WSL (Windows Subsystem for Linux). Install with: wsl --install

Full setup details: Installation Guide

Which method?

Scenario Use Why
First time CLI Stable, full-featured
Team rollout CLI Same command everywhere
Early adopter Plugin Zero dependencies, instant (beta)
Contributing CLI Dev scripts, internals access
Already on CLI Either Both coexist safely

Use (inside Claude Code, after reopening it)

/nw:discuss "user login with email and password"   # Requirements
/nw:design --architecture=hexagonal                 # Architecture
/nw:distill "user-login"                            # Acceptance tests
/nw:deliver                                         # TDD implementation

Four commands. Four human checkpoints. One working feature.

Full walkthrough: Your First Feature

Staying Updated

nWave checks for new versions when you open Claude Code. When available, you'll see a note in Claude's context with version details and changes.

Plugin method:

/plugin marketplace update nwave-marketplace

Or enable auto-updates in Claude Code plugin settings.

CLI method:

pipx upgrade nwave-ai
nwave-ai install

Control check frequency via update_check.frequency in ~/.nwave/des-config.json: daily, weekly, every_session, or never.

Uninstalling

Plugin method:

/plugin uninstall nw

CLI method:

nwave-ai uninstall              # Remove agents, commands, config, DES hooks
pipx uninstall nwave-ai        # Remove the Python package

Both methods remove agents, commands, and configuration from ~/.claude/. Your project files are unaffected.

Token Efficiency — Scale Quality to Stakes

nWave enforces proven engineering practices (TDD, peer review, mutation testing) at every step. Use /nw:rigor to adjust the depth of quality practices to match your task's risk level. A config tweak needs less rigor than a security-critical feature.

/nw:rigor                    # Interactive: compare profiles
/nw:rigor lean               # Quick switch to lean mode
/nw:rigor custom             # Build your own combination
Profile Agent Reviewer TDD Mutation Cost Use When
lean haiku none RED→GREEN no lowest Spikes, config, docs
standard sonnet haiku full 5-phase no moderate Most features
thorough opus sonnet full 5-phase no higher Critical features
exhaustive opus opus full 5-phase ≥80% kill highest Production core
custom you choose you choose you choose you choose varies Exact combo needed

Picked once, persists across sessions. Every /nw:deliver, /nw:design, /nw:review respects your choice. Need to mix profiles? /nw:rigor custom walks through each setting.

/nw:rigor lean        # prototype fast
/nw:deliver           # haiku crafter, no review, RED→GREEN only
/nw:rigor standard    # ready to ship — bump up
/nw:deliver           # sonnet crafter, haiku reviewer, full TDD

Understanding DES Messages

DES is nWave's quality enforcement layer — it monitors every agent Task invocation during feature delivery to prevent unbounded execution, enforce TDD discipline, and protect accidental edits. Most DES messages are normal enforcement, not errors. They appear when agents skip required safety checks or when your code contains patterns that look like step execution.

DES also runs automatic housekeeping at every session start: it removes audit logs beyond the retention window, cleans up signal files left by crashed sessions, and rotates the skill-loading log when it grows too large. This happens silently in the background and never blocks your session.

Message What It Means What To Do
MISSING_MAX_TURNS Task invocation forgot to set max_turns parameter. Add max_turns=30 (or appropriate value) to the Task call. Recommended: 15 (quick), 25 (background), 30 (standard), 35 (research).
DES_MARKERS_MISSING Task prompt mentions a step ID (01-01 pattern) but lacks DES markers. Either: add DES markers for step execution, OR add <!-- DES-ENFORCEMENT : exempt --> comment if it's not actually step work.
Source write blocked You tried to edit a file during active /nw:deliver outside a DES task. Edit requests must go through the active deliver session. If you need to make changes, finalize the current session first.
TDD phase incomplete Sub-agent returned without finishing all required TDD phases. Re-dispatch the same agent to complete missing phases (typically COMMIT or refactoring steps).
nWave update available SessionStart detected a newer version available. Optional. Run pipx upgrade nwave-ai && nwave-ai install when ready to upgrade, or dismiss and continue working.
False positive blocks Your prompt accidentally matches step-ID pattern (e.g., dates like "2026-02-09"). Add <!-- DES-ENFORCEMENT : exempt --> comment to exempt the Task from step-ID enforcement.

These messages protect code quality but never prevent your work — they guide you toward the safe path.

How It Works

  machine        human         machine        human         machine
    │              │              │              │              │
    ▼              ▼              ▼              ▼              ▼
  Agent ──→ Documentation ──→ Review ──→ Decision ──→ Agent ──→ ...
 generates    artifacts      validates   approves    continues

Each wave produces artifacts that you review before the next wave begins. The machine never runs unsupervised end-to-end.

The full workflow has six waves. Use all six for greenfield projects, or jump straight to /nw:deliver for brownfield work.

Wave Command Agent Produces
DISCOVER /nw:discover product-discoverer Market validation
DISCUSS /nw:discuss product-owner Requirements
DESIGN /nw:design solution-architect Architecture + ADRs
DEVOPS /nw:devops platform-architect Infrastructure readiness
DISTILL /nw:distill acceptance-designer Given-When-Then tests
DELIVER /nw:deliver software-crafter Working implementation

23 agents total: 6 wave agents, 6 cross-wave specialists, 11 peer reviewers. Full list: Commands Reference

Documentation

Getting Started

Guides & Reference

Community

Privacy

nWave does not collect user data. See Privacy Policy for details.

License

MIT — see LICENSE for details.

Project details


Release history Release notifications | RSS feed

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

nwave_ai-1.7.4rc1.tar.gz (2.7 MB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

nwave_ai-1.7.4rc1-py3-none-any.whl (1.0 MB view details)

Uploaded Python 3

File details

Details for the file nwave_ai-1.7.4rc1.tar.gz.

File metadata

  • Download URL: nwave_ai-1.7.4rc1.tar.gz
  • Upload date:
  • Size: 2.7 MB
  • Tags: Source
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/6.1.0 CPython/3.13.7

File hashes

Hashes for nwave_ai-1.7.4rc1.tar.gz
Algorithm Hash digest
SHA256 8f2fc39144d86e4a4ce1df2eade65b30a3acc956d04979ad5c8c329defea7a15
MD5 8de6773dedc954d2c18fa7f7ff8a5542
BLAKE2b-256 0341bfbbfb1d60bcc00cfb3004450c1df3984f249355da22c211820698442f8f

See more details on using hashes here.

Provenance

The following attestation bundles were made for nwave_ai-1.7.4rc1.tar.gz:

Publisher: release-rc.yml on nWave-ai/nwave-dev

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file nwave_ai-1.7.4rc1-py3-none-any.whl.

File metadata

  • Download URL: nwave_ai-1.7.4rc1-py3-none-any.whl
  • Upload date:
  • Size: 1.0 MB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/6.1.0 CPython/3.13.7

File hashes

Hashes for nwave_ai-1.7.4rc1-py3-none-any.whl
Algorithm Hash digest
SHA256 20bc33d79e1fc225b46b996db21261a6f95362203547b449fef3cd8641a0da7d
MD5 b57f9e54d3f81934b96a99710ca59d14
BLAKE2b-256 851fd1c6804924fa4fa3a058bdde3a6419e49fa9754279d8c7211f81252fe113

See more details on using hashes here.

Provenance

The following attestation bundles were made for nwave_ai-1.7.4rc1-py3-none-any.whl:

Publisher: release-rc.yml on nWave-ai/nwave-dev

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page