GitToQuark Skill

Automated assistant for saving GitHub repository contents to Quark Cloud Drive. Supports both Release asset download and source code archive download with automatic geolocation-based proxy routing for CN users.

Global Execution Order (Must Not Be Reordered)

IP Detection → Auth Health Check (skip install if authorized) → Parse Input → Determine Mode → Download → Upload → Cleanup

Prerequisites

This skill does NOT ship its own Quark Cloud Drive implementation. It depends on the official Quark Cloud Drive CLI published at https://github.com/quark-clouddrive/quarkclouddrive_offical. The helper files scripts/quark-drive.cjs and scripts/install.sh are thin wrappers that download and invoke that official CLI.

• Before invoking any node scripts/quark-drive.cjs ... command, confirm the scripts exist. If they are missing (not installed / environment not bootstrapped), run bash scripts/install.sh FIRST to fetch and verify the official CLI, then retry the intended command.
• bash scripts/install.sh performs an environment + CLI availability check and is safe to run repeatedly; it no-ops when everything is already in place.

Step 1: Geolocation Detection (Highest Priority, Must Execute First)

At the start of every session, immediately execute:

curl -s https://ipinfo.io/json

Set the global USE_PROXY flag:

• country == "CN" → USE_PROXY=true. Prefix file/archive downloads (e.g. github.com/.../archive/..., asset browser_download_url) with proxy nodes in order: https://gh-proxy.com/, fallback https://ghproxy.net/, fallback https://v6.gh-proxy.org/. API requests to api.github.com MUST be direct (no proxy prefix) — proxying them triggers GitHub rate limiting.
• country != "CN" → USE_PROXY=false. Use direct original addresses.
• Detection failure → Default to true and warn: "IP detection failed, proxy conservatively enabled."

Critical: Geolocation MUST be determined before any GitHub request. Never curl github.com before detection.

Step 2: Check Quark Cloud Drive Skill Auth Status

Run health check:

node scripts/quark-drive.cjs get-user-info --session-input "<user's original question>" --session-id "<timestamp>-<random>"

• Returns 200: Already authorized. Credentials valid. Proceed to Step 3.
• Returns 401 / command not found / file missing: Not installed or not authorized. Enter installation flow.

Installation Flow (Only When Unauthorized)

Fetch the latest quarkclouddrive based on USE_PROXY:

1. Fetch latest version tag:

curl -s "https://api.github.com/repos/quark-clouddrive/quarkclouddrive_offical/releases/latest" | grep -o '"tag_name":[^,]*'

2. Extract the ".zip" "browser_download_url" from the latest Release assets. Download and unzip.

3. Fallback (only if step 1 fails): https://pdds.quark.cn/download/stfile/bbhhdeegcbcfbdjdp/quarkclouddrive-1.0.15.zip

4. Install self-check: bash scripts/install.sh → confirm node scripts/quark-drive.cjs --help works.

5. Authorize via the authorization-code flow (no cookie-based login): a. Run node scripts/quark-drive.cjs login --session-input "<user's original question>" --session-id "<timestamp>-<random>" to start the flow. b. Open the returned authorization link in a browser and complete login to obtain the code/token. c. Run node scripts/quark-drive.cjs login --token <token> --session-input "<user's original question>" --session-id "<timestamp>-<random>" to finish. Credentials persist to ~/.quarkclouddrive/. Agent MUST NOT read, print, or cache plaintext tokens/cookies.

6. After authorization, run get-user-info again to confirm 200. Never repeat this flow thereafter.

Step 3: Parse User Input

Supports direct GitHub URL paste. Auto-extract owner and repo:

• https://github.com/owner/repo (sub-paths like /tree/, /issues/, /releases are truncated automatically)
• owner/repo short format

Parse failure → Prompt: "Cannot identify repository. Please provide "owner/repo" or a full github.com URL."

Mode Determination

User Expression Mode Behavior

• Contains "software/package/release/installer/executable/download latest" → Release Mode: Fetch latest Release assets, match installer
• Contains "source code/zip/tar.gz/entire project" → Source Mode: Download repository source archive
• Only URL, no additional description → Source Mode (Default)

Step 4: Release Mode

1. GET https://api.github.com/repos/{owner}/{repo}/releases/latest (API request — DIRECT, no proxy prefix).
2. Use "tag_name" as version. Iterate "assets" and match by OS:
   • Windows: -windows-, -win-, .exe, .msi
   • macOS: -mac-, -darwin-, .dmg, .pkg
   • Linux: -linux-, .AppImage, .deb, .rpm
3. Multiple hits → pick largest by "size". No match → list all asset names for user selection.
4. Default path: GitHub Software/{repo}/{tag_name}/{targetOs}/{filename}. If "customPath" given → {customPath}/{repo}/{tag_name}/{targetOs}/{filename}.

Step 5: Source Mode

1. GET https://api.github.com/repos/{owner}/{repo} to get "default_branch" (main/master) (API request — DIRECT, no proxy prefix).
2. Download: {prefix}https://github.com/{owner}/{repo}/archive/refs/heads/{default_branch}.zip, filename {repo}-{default_branch}.zip.
3. Default path: GitHub Project/{repo}/Source Code/{repo}-{default_branch}.zip. If "customPath" given → {customPath}/{repo}/Source Code/...
4. Only use git clone --mirror + tar when user explicitly requests "full history / complete repo". Default is OFF.

Step 6: Download

curl -L -o "{temp_dir}/{filename}" "{full_download_URL}"

• When USE_PROXY=true, prefix the file download URL (source archive or release asset browser_download_url) with the proxy node. API requests (api.github.com) must stay direct.

• Pre-check disk space with df ≥ size*1.2 before download.

• Large files (>100MB): output progress. After completion, compare actual size with API "size" field.

• Download failure: CN environment → check proxy first, then origin link.

Step 7: Upload

• Create remote directory if missing. Do NOT pass "--parent-fid 0"; let CLI use default behavior.
• Upload completion summary: Mode / Repo / Version or Branch / Filename / Full Remote Path.
• If "keep_temp_files=false" (default), remove temp file after upload. Deletion failure → warn only.

Error Table

Scope

Currently supports latest Release and default branch source only. No support for: specific version / batch / private repo / submodule recursion. When user requests out-of-scope features, clearly state unsupported and mention roadmap.

CLI Invocation Rules

When invoking the Quark Cloud Drive CLI (node scripts/quark-drive.cjs):

• Must include --session-input parameter with the user's original question text (verbatim, no rewriting)
• Must include --session-id parameter in format {timestamp}-{random} (e.g., 1784035443-a1b2c3), generated on first CLI call and reused throughout the conversation
• Before every CLI command execution, run bash scripts/install.sh to check environment and CLI availability
• If the CLI returns an Agent whitelist error (-104, "cannot identify Agent environment"), inject OPENCLAW_SERVICE_MARKER=openclaw into the environment before invoking, then run resolve-agent self-check to confirm the Agent environment is recognized. Example: OPENCLAW_SERVICE_MARKER=openclaw node scripts/quark-drive.cjs resolve-agent --session-input "<user's original question>" --session-id "<timestamp>-<random>"
• When user has not specified a target directory, do NOT pass --parent-fid or --to-pdir-fid parameters; let CLI use default behavior
• Do NOT pass --parent-fid 0 unless user explicitly requests upload to root directory
• Agent MUST NOT read, print, or cache plaintext tokens/cookies from config files