본문으로 건너뛰기
cccskills
무료GitHub에 공개

desktop-brand-builder

Build a custom branded Qwen Code desktop package from its Tauri shell using a brand identifier and local logo. Apply app naming, icons, and startup screen changes, then report generated installers and their verification hashes.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

원문원문 설명 보기

Generate a branded Qwen Code desktop package from the Tauri desktop shell using a minimal brandId and logo. Use when the user wants a custom, white-label, or rebranded desktop client, installer, DMG/EXE/AppImage/deb, or one-click brand build on top of packages/desktop.

설치 방법 보기

이럴 때 유용합니다

  • Create a desktop client with a company name and logo
  • Build an installer for a supported operating system
  • Configure a separate update feed for a branded app
  • Verify generated packages and record their hashes

포함된 파일(2)

  • SKILL.md8.4 KB
  • scripts/brand-create.mjs13.2 KB

SKILL.md(원문)

설치하기 전에 에이전트에게 전달되는 지시 내용을 확인할 수 있습니다.

Desktop Brand Builder (Tauri shell)

Goal

Create a branded desktop package from packages/desktop with the least user input possible. The user should usually provide only:

brandId: acme-ai
logo: /absolute/path/to/logo.png
website: https://acme.ai

website is optional. Do not ask for app name, app id, artifact name, copyright, or updater endpoints unless the user explicitly asks to override them.

This skill replaces the Electron-era brand builder that lived in the older, since-deleted Electron packages/desktop (removed in PR 9085, before this Tauri shell took over that path). The Tauri shell is the only desktop implementation now; branding hooks are src-tauri/tauri.conf.json, src-tauri/icons/, and the bootstrap/ startup UI.

Input Rules

Required fields:

  • brandId: must match ^[a-z][a-z0-9-]*$
  • logo: local file path; must exist; .png recommended (square, >= 1024px)

Optional overrides:

  • website
  • appName
  • appId (Tauri bundle identifier)
  • artifactPrefix
  • updaterEndpoints (JSON array; empty array disables in-app updates)
  • updaterPubkey (base64 public key; required when updaterEndpoints is non-empty — must match the TAURI_SIGNING_PRIVATE_KEY used to sign your updater artifacts)
  • target: mac, win, linux, or all

If required input is missing, ask once:

请提供:
brandId: 例如 acme-ai,只能小写字母、数字、短横线
logo: 本地 logo 文件路径(建议 1024x1024 PNG)
website: 可选

Once the required fields are present, proceed without a confirmation step.

Derived Defaults

Infer missing values deterministically:

  • appName: title-case the hyphen-separated brandId; acme-ai becomes Acme AI
  • artifactPrefix: title-case the hyphen-separated brandId and join with hyphens; acme-ai becomes Acme-AI
  • appId: if website has a valid host, reverse the host labels and append .desktop; https://acme.ai becomes ai.acme.desktop
  • fallback appId: app.<brandId>.desktop
  • updaterEndpoints: empty by default. A branded build must never poll the official Qwen Code updater feed, and the official feed must never update a branded build. Only set endpoints when the user supplies their own feed.

Workflow

Work in an isolated build clone so the working repository stays clean:

BUILD_ROOT="$PWD/brand-builds/<brandId>-<timestamp>"
mkdir -p "$BUILD_ROOT"
git clone --branch main --single-branch \
  https://github.com/QwenLM/qwen-code.git \
  "$BUILD_ROOT/qwen-code"
cd "$BUILD_ROOT/qwen-code"
git checkout -B brand-<brandId> origin/main

If the clone or checkout fails, stop and report the failure. Do not continue as if brand-<brandId> was created.

Create a temporary brand.json in the build directory:

{
  "brandId": "acme-ai",
  "logo": "/absolute/path/to/logo.png",
  "website": "https://acme.ai",
  "appName": "Acme AI",
  "appId": "ai.acme.desktop",
  "artifactPrefix": "Acme-AI",
  "updaterEndpoints": [],
  "updaterPubkey": ""
}

Install dependencies. The brand script itself only needs the desktop package's own node_modules, but npm run build:runtime shells out to the repo root (which uses cross-env and other root devDependencies), so the root install is also required before packaging:

# Root dependencies (needed by build:runtime → cross-env, esbuild, etc.)
npm install

# Desktop package dependencies
cd packages/desktop
npm install --workspaces=false
cd ../..

Then run this skill's bundled brand creation script with plain Node (the script has no dependencies beyond Node >= 18):

node packages/desktop/.agents/skills/desktop-brand-builder/scripts/brand-create.mjs \
  --shell-root /absolute/path/to/qwen-code/packages/desktop \
  --config /absolute/path/to/brand.json

The agent should not hand-edit tauri.conf.json, icon files, or bootstrap brand strings when this bundled script is available. The bundled script is the source of truth for patching config and generating resources.

What the script does:

  1. Patches src-tauri/tauri.conf.json: productName, identifier, bundle.shortDescription, and plugins.updater.endpoints. When updaterEndpoints is empty it also clears bundle.createUpdaterArtifacts and blanks the official plugins.updater.pubkey (set to empty string rather than deleted, because the updater plugin requires the field); a brand supplying its own feed must supply its own pubkey.
  2. Regenerates the full icon set from the logo via npx --yes @tauri-apps/cli icon <logo> (falls back to a warning if the CLI cannot run; in that case copy the logo over src-tauri/icons/icon.png manually and tell the user the remaining sizes are stale).
  3. Patches the bootstrap UI: page title, brand heading, startup strings in bootstrap/index.html and bootstrap/bootstrap.js, and replaces bootstrap/qwen-code-logo.svg usage with the brand logo.

Package with the current host target unless the user requested a target:

cd packages/desktop
npm run build:runtime --workspaces=false
npx tauri build            # current platform

Cross-compile: build:runtime bundles the Node runtime for the platform indicated by QWEN_DESKTOP_TARGET (defaults to the host). When targeting a different platform you must re-run build:runtime with the env var set before each tauri build --target, otherwise the packaged artifact contains a wrong-arch Node binary and fails at launch with an exec format error:

# Cross-compile: set QWEN_DESKTOP_TARGET and re-run build:runtime per target
QWEN_DESKTOP_TARGET=aarch64-apple-darwin npm run build:runtime --workspaces=false
npx tauri build --target aarch64-apple-darwin   # explicit macOS arm64

For target: all, iterate build:runtime → tauri build per target; run only targets supported by the current machine or CI environment. Do not claim cross-platform artifacts were produced unless the files exist. Artifacts land under packages/desktop/src-tauri/target/release/bundle/ for the host target, or src-tauri/target/<triple>/release/bundle/ when --target <triple> is used.

Signing and Updates

Branded builds are unsigned by default. The upstream release pipeline's signing secrets (Apple, Windows) and updater private key belong to the official Qwen Code releases only. For a brand that needs signed releases or in-app updates, set up separate credentials and a separate updater feed; do not reuse the upstream ones.

To generate a signing key pair for your updater feed:

npx @tauri-apps/cli signer generate -w ~/.tauri/my-brand.key
# The .key file is the private key (set as TAURI_SIGNING_PRIVATE_KEY in
# your build CI). The corresponding .pub file contains the base64 public
# key — paste it into brand.json as updaterPubkey.

Validation

After packaging:

  1. Confirm the expected artifact exists under packages/desktop/src-tauri/target/release/bundle/ (or src-tauri/target/<triple>/release/bundle/ for cross-compile targets) (dmg/, nsis/, appimage/, or deb/).
  2. Compute sha256sum or shasum -a 256 for each artifact.
  3. On macOS, run hdiutil verify for generated DMG files.
  4. Report the artifact path, SHA-256, app name, app id, and build directory.

Failure Handling

  • Invalid brandId: show the regex and ask for a corrected value.
  • Missing logo: ask for a valid local path.
  • Missing bundled script: report that packages/desktop/.agents/skills/desktop-brand-builder/scripts/brand-create.mjs is missing, and include the expected command.
  • Already-branded shell-root: the script refuses to run when productName is no longer the default (Qwen Code Desktop). Start from a fresh clone — do not re-run the script in an already-patched tree.
  • Build failure: preserve the build directory, return the last useful error lines, and include the full log path or command that produced the failure.

Do not delete the build directory on failure. Never re-run brand-create in the same clone — the script is single-use. The "preserve the build directory" guidance is for post-mortem debugging, not for retrying the brand step. If the brand config was wrong, discard the clone and start fresh.

리뷰

아직 리뷰가 없습니다. 사용해 본 소감을 남겨 주세요.

댓글

로그인 없이 쓸 수 있습니다. 사용 후기, 질문, 작동하지 않은 환경 등을 편하게 남겨 주세요.

아직 댓글이 없습니다. 첫 댓글을 남겨 보세요.

보내면 바로 공개됩니다. 이메일 주소나 전화번호 같은 개인 정보는 쓰지 마세요. URL은 2개까지입니다.

0 / 1,000자

이 양식은 reCAPTCHA로 보호되며 Google 개인정보처리방침 및 서비스 약관이 적용됩니다.

같은 저장소의 스킬

Write clear briefs for subagents and conversation forks, agents that inherit conversation history. Specify purpose, known findings, scope, and expected output while respecting each agent's existing role and separating research requests from implementation.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

  • Writing a focused research brief for a subagent
  • Specifying files and constraints for implementation
  • Defining scope for an agent inheriting conversation history
QwenLM/qwen-code2.9만2026년 10월 19일 업데이트

Compare a feature implemented in Qwen Code with Codex or Claude Code under the same scenario. Capture communication and terminal records, inspect tool definitions and outputs, and guide fixes toward matching observable behavior.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

  • Checking a reproduced feature against its reference agent
  • Comparing tool definitions and required arguments
  • Investigating differences in terminal output
QwenLM/qwen-code2.9만2026년 10월 19일 업데이트

Reproduce a Codex or Claude Code feature in Qwen Code or another command-line agent. Capture requests, terminal output, and state changes, then use that evidence to implement focused behavior and add a verification path.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

  • Reproduce an existing agent's slash command
  • Inspect request formats and tool definitions
  • Compare terminal output and saved state
QwenLM/qwen-code2.9만2026년 10월 19일 업데이트

autofix

무료

Review local uncommitted changes, verify findings, and repeat focused repairs until the review converges or a blocker remains. Also supports Qwen Code issue fixes and review responses within workflows run through GitHub Actions.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

  • Review and repair local uncommitted changes
  • Develop a focused fix for a selected issue
  • Verify and address pull request feedback
QwenLM/qwen-code2.9만2026년 10월 19일 업데이트

batch

무료

Apply the same operation across multiple files using parallel worker agents. The skill finds matching files, divides the work, and reports successes, failures, and skipped files, with a preview option before edits.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

  • Add license headers across source files
  • Convert JavaScript files to TypeScript
  • Fix the same lint issue in multiple files
QwenLM/qwen-code2.9만2026년 10월 19일 업데이트

batch-api

무료

Prepare shared instructions for translating, rewriting, or extracting information from many independent files, then submit a plan through the DashScope Batch API. Processing runs asynchronously at half the realtime list price and delivers new files later.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

  • Translate many documents using shared terminology
  • Rewrite independent files into a consistent format
  • Extract structured data from a collection of files
QwenLM/qwen-code2.9만2026년 10월 19일 업데이트

QwenLM의 스킬 모두 보기

이 스킬의 문제 신고