mcpField
Z
zeabur-file

zeabur-file

github.com/zeaburDevelopment39118 installsUpdated 7w ago
Repository

About zeabur-file

zeabur-file is an agent skill for retrieving uploaded project files with the Zeabur CLI and examining their structure and configuration. It guides agents through an ordered inspection process before summarizing findings and suggesting deployment with zeabur-deploy.

Key features

  1. 01Pull uploaded files with the Zeabur CLI
  2. 02Skip binary files during downloads
  3. 03Preserve directory structure
  4. 04Inspect source files and project configuration
  5. 05Recommend an ordered analysis workflow

Use cases

  1. 01Inspect a project provided through an upload_id
  2. 02Review source files and build configuration before deployment

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/zeabur/agent-skills --skill zeabur-file

After installing, type/zeabur-filein Claude Code to confirm it loaded.

Details

From the project README

Zeabur File

Always use npx zeabur@latest to invoke Zeabur CLI. Never use zeabur directly or any other installation method. If npx is not available, install Node.js first.

Trigger

When the user's message contains <UploadedFile>upload_id</UploadedFile>, extract the upload_id from inside the tag.

Workflow

# 1. Pull the uploaded project to a local directory
npx zeabur@latest file pull <upload_id> /tmp/project -i=false

# 2. Explore with standard bash tools
ls /tmp/project
cat /tmp/project/package.json
find /tmp/project -name "*.ts" -o -name "*.js" | head -20

Command

# Download all files from an upload to a local directory
npx zeabur@latest file pull <upload_id> [target-dir] -i=false
  • Default target directory is . (current directory)
  • Binary files (images, fonts, archives, etc.) are skipped automatically
  • Directory structure is preserved

Recommended Analysis Order

  1. Pull the project: npx zeabur@latest file pull <upload_id> /tmp/project -i=false
  2. List structure: ls -la /tmp/project and find /tmp/project -type f
  3. Read config: cat /tmp/project/package.json (or go.mod, requirements.txt, Cargo.toml)
  4. Read Dockerfile if present
  5. Read entry point files (src/main.*, index.*, app.*)
  6. Summarize findings and suggest deployment with the zeabur-deploy skill

Tips

TipDetails
Binary files are skippedImages, fonts, archives won't be downloaded
Use a temp directoryPull to /tmp/project to keep the workspace clean
Use standard toolsAfter pulling, grep, find, wc, cat all work normally

Comments

Markdown supported