Queues a deferred self-update of nwave-ai. Writes a PendingUpdateFlag that the SessionStart hook replays on the next Claude Code launch, so the current session is not interrupted. Falls back to manual instructions when the package manager cannot be detected.
npx skills add https://github.com/nWave-ai/nWave --skill nw-update
Wave: CROSS_WAVE | Agent: Main Instance (self) | Command: /nw-update
/nw-update records a PendingUpdateFlag on disk. The actual upgrade is performed
by the SessionStart hook on the next Claude Code launch, so the user can keep
working in the current session and apply the new version whenever they restart.
You (the main Claude instance) run this command directly — no subagent delegation.
Do NOT attempt to upgrade nwave-ai in the current session.
$ARGUMENTS), treat it as the target versionafter stripping a leading v.
from the global config (update_check.latest_available):
python3 -c "
import sys, os
sys.path.insert(0, os.path.expanduser('~/.claude/lib/python'))
from des.adapters.driven.config.des_config import DESConfig
print(DESConfig().update_check_latest_available or '')"
target version explicitly. Do not guess.
Invoke the detector and resolve the PM binary absolute path in a single Bash call.
This runs under whatever python3 the shell provides, which is NOT the
interpreter nwave-ai was installed under — so sys.executable is the wrong
anchor for detection. Use resolve_nwave_pm, which prefers the value the
installer recorded at install time (install.package_manager in the global
config) and only falls back to live path detection:
python3 -c "
import sys, os
sys.path.insert(0, os.path.expanduser('~/.claude/lib/python'))
import shutil
from pathlib import Path
from des.adapters.driven.config.des_config import DESConfig
from des.adapters.driven.package_managers.package_manager_detector import (
resolve_nwave_pm,
)
pm = resolve_nwave_pm(DESConfig().installed_package_manager, Path(sys.executable))
binary = shutil.which(pm) if pm in ('pipx', 'uv') else ''
print(f'{pm}|{binary or \"\"}')"
Parse the output as pm|binary_abspath.
If pm == "unknown" OR pm == "pip" OR binary_abspath is empty:
(detect_pm returns pip only when forced via NWAVE_INSTALLER=pip; pip
installs are not auto-updatable through the deferred-upgrade flow, so they take
the manual path alongside unknown.)
PendingUpdateService.request_update.nwave-ai was not installed via an auto-updatable package manager.
Please upgrade manually:
uv tool install nwave-ai@latest && nwave-ai install
— or —
pipx upgrade nwave-ai && nwave-ai install
— or, if you installed with pip —
pip install -U nwave-ai && nwave-ai install
Then restart Claude Code.
Call the driving port PendingUpdateService.request_update:
Reuse the pm and binary already resolved in Step 2 — do NOT re-detect from
sys.executable here (that interpreter is not the install anchor). Construct
DESConfig() directly; there is no DESConfig.load().
python3 -c "
import sys, os
sys.path.insert(0, os.path.expanduser('~/.claude/lib/python'))
from des.application.pending_update_service import PendingUpdateService
from des.adapters.driven.config.des_config import DESConfig
pm = '${PM}'
binary = '${PM_BINARY}'
target = '${TARGET_VERSION}'
config = DESConfig()
# PM adapter is only used by apply(); request_update() does not invoke it.
svc = PendingUpdateService(config=config, pm=None) # type: ignore[arg-type]
svc.request_update(pm=pm, pm_binary_abspath=binary, target_version=target)
print(f'queued:{pm}:{target}')"
Substitute ${PM} and ${PM_BINARY} with the Step 2 values, and
${TARGET_VERSION} with the Step 1 value (no leading v).
On success, print:
Update to vX.Y.Z queued. It will be applied the next time you restart Claude Code.
Package manager: <pipx|uv> (<binary_abspath>)
| Condition | Response |
|-----------|----------|
| No target version provided and none discovered | Ask the user for an explicit version |
| PM detection returns unknown | Print manual fallback (Step 3), do not write flag |
| shutil.which(pm) returns empty | Same as unknown — print fallback |
| request_update raises | Surface the exception, do not claim success |
This command is synchronous and fast (< 1s). Do not create a task list; print
progress inline.
/nw-update
Reads update_check.latest_available = "1.4.2", detects pipx at
/home/alex/.local/bin/pipx, queues flag, prints:
Update to v1.4.2 queued. It will be applied the next time you restart Claude Code.
/nw-update 1.5.0
Detects uv at /home/alex/.cargo/bin/uv, queues flag for 1.5.0.
/nw-update 1.4.2
Detector returns unknown (e.g. pip venv install). Prints manual fallback
instructions and exits without writing the flag.
Guide for creating high-quality MCP (Model Context Protocol) servers that enable LLMs to interact with external services through well-designed tools. Use when building MCP servers to integrate external APIs or services, whether in Python (FastMCP) or Node/TypeScript (MCP SDK).
Automatically creates user-facing changelogs from git commits by analyzing commit history, categorizing changes, and transforming technical commits into clear, customer-friendly release notes. Turns hours of manual changelog writing into minutes of automated generation.
Use when implementation is complete, all tests pass, and you need to decide how to integrate the work - guides completion of development work by presenting structured options for merge, PR, or cleanup
Guide for creating high-quality MCP (Model Context Protocol) servers that enable LLMs to interact with external services through well-designed tools. Use when building MCP servers to integrate external APIs or services, whether in Python (FastMCP) or Node/TypeScript (MCP SDK).
React Native and Expo best practices for building performant mobile apps. Use when building React Native components, optimizing list performance, implementing animations, or working with native modules. Triggers on tasks involving React Native, Expo, mobile performance, or native platform APIs.
React and Next.js performance optimization guidelines from Vercel Engineering. This skill should be used when writing, reviewing, or refactoring React/Next.js code to ensure optimal performance patterns. Triggers on tasks involving React components, Next.js pages, data fetching, bundle optimization, or performance improvements.
Next.js best practices - file conventions, RSC boundaries, data patterns, async APIs, metadata, error handling, route handlers, image/font optimization, bundling
Use when starting feature work that needs isolation from current workspace or before executing implementation plans - creates isolated git worktrees with smart directory selection and safety verification
Take nwave-ai/nw-update from the repository into ~/.claude/skills for personal
use, or into .claude/skills inside a project.
The agent identifies a skill by the name field in its header. Two skills with the
same name cannot sit side by side — one of them will be ignored.
The instructions reference pip.
Without those the skill loads but fails at the first command.