GitHub Memory Sync

将 OpenClaw 的完整工作空间配置（SOUL.md, IDENTITY.md, USER.md, MEMORY.md, TOOLS.md, memory/*等）同步到 GitHub 进行备份和版本控制，支持跨服务器迁移。Use when user mentions GitHub backup, sync m...

Security Scan

The skill mostly does what it says (backup OpenClaw workspace to GitHub) but has several inconsistencies and insecure defaults that could accidentally expose sensitive credentials or workspace secrets.

Purpose & Capability

The skill's stated purpose (backup/sync OpenClaw workspace to GitHub) aligns with the included scripts and docs. However there are manifest inconsistencies: the top-level registry metadata claims no required env vars, while clawhub.yaml and SKILL.md require GITHUBTOKEN and GITHUB_REPO. The included LICENSE text (GPLv3) conflicts with metadata saying MIT. These mismatches indicate sloppy packaging and deserve attention before trusting the skill.

Instruction Scope

SKILL.md and the scripts explicitly copy highly sensitive files (SOUL.md, IDENTITY.md, USER.md, MEMORY.md, memory/*, skills/, projects/agents, BOOTSTRAP.md). Although SKILL.md warns not to upload openclaw.json channel credentials, the scripts do not explicitly exclude files like openclaw.json or other secret-containing files. The docs even demonstrate adding tokens inline in cron lines (which would store tokens in crontab). This combination creates a real risk of accidental secrets exfiltration.

Install Mechanism

No external downloads or package installs are performed; this is an instruction-only skill with two shell scripts. That reduces supply-chain risk. The scripts are executable and will be run on the host, so local script behavior is the primary risk rather than remote fetches.

Credentials

The skill requires a GitHub PAT (GITHUBTOKEN) and repository name — reasonable for GitHub sync. But the scripts embed the token into HTTPS clone/push URLs (REPO_URL="https://${GITHUB_TOKEN}@github.com/${GITHUB_REPO}.git"), which can cause credentials to be saved in git metadata or appear in process contexts. The cron examples and CRON.md show placing tokens directly in crontab or environment files, raising further exposure risk. Defaults in cron-backup.sh (a hardcoded GITHUB_REPO owner) are suspicious and may cause misconfiguration.

Persistence & Privilege

The skill does not request always:true and is user-invocable only. However the documentation encourages adding a cron job for automated backups and shows patterns that would persist credentials (crontab/environment) on the host. Persisting scheduled execution combined with insecure token handling increases blast radius; treat automatic installation of cron entries with caution.

What to consider before installing

This skill's functionality matches its description, but it contains insecure defaults and packaging mistakes you should address before using it on sensitive data. Key actions to consider before installing or running: - Do not paste a long-lived PAT into crontab or /etc/environment. Use short-lived tokens or a deploy key with minimal scope when possible, and rotate tokens after testing. - The scripts build an HTTPS URL with the token embedded. That can leave tokens in .git/config or process contexts. Prefer using a credential helper, the GitHub CLI (gh auth), or SSH deploy keys instead of embedding tokens in URLs. Remove credentials from .git/config after pushes. - The scripts copy many files and do not explicitly exclude openclaw.json or other config files; add explicit excludes (e.g., a .gitignore or modify copy_workspace_files to skip openclaw.json, any secrets/ or tokens files) and verify the repo contents before pushing. - Avoid backing up sensitive channel credentials; manually review files to be synced or run the script with --memory-only if you only need memory files. - Fix manifest mismatches (registry env metadata vs clawhub.yaml, and license mismatch) or request a corrected package from the author before trusting it. - If you must use automation, prefer a systemd timer running a wrapper that loads credentials from a protected file with restricted permissions (not crontab), and ensure /var/log/openclaw-memory-sync.log is set with appropriate permissions and log-rotation. If you cannot verify or modify the scripts, treat this skill as high-risk for accidental credential exposure and do not use it with production secrets or channel tokens.

GitHub Memory Sync 技能

📝 将 OpenClaw 的 完整工作空间配置 同步到 GitHub 仓库进行备份和版本控制，支持跨服务器迁移。

同步范围

核心记忆文件（必须同步）

文件	路径	说明	敏感度
SOUL.md	/	AI 人格定义	🔒 高
IDENTITY.md	/	AI 身份定义（名字、emoji 等）	🔒 高
USER.md	/	用户信息	🔒 高
MEMORY.md	/	长期记忆	🔒 高
TOOLS.md	/	工具配置（SSH、摄像头等）	🔒 高
HEARTBEAT.md	/	心跳任务配置	🟡 中
memory/*.md	memory/	日常记忆文件	🔒 高

可选配置文件

文件	路径	说明	建议
AGENTS.md	/	工作空间指南	✅ 推荐
BOOTSTRAP.md	/	初始化脚本（如有）	⚪ 可选
skills/	skills/	自定义技能	✅ 推荐
avatars/	avatars/	头像图片	⚪ 可选

排除文件（不同步）

• .git/ - Git 元数据
• node_modules/ - 依赖包
• *.log - 日志文件
• *.tmp, *.bak - 临时文件
• sessions/ - 会话数据（可能很大）

功能特性

1. 📤 完整备份 - 将所有记忆和配置文件推送到 GitHub
2. 📥 一键恢复 - 从 GitHub 拉取配置到新服务器
3. 📊 查看状态 - 检查本地和远程的差异
4. 📋 列出文件 - 显示所有同步的文件
5. 🔧 初始化仓库 - 首次设置 GitHub 仓库连接
6. 🔄 增量同步 - 只同步变化的文件
7. 📦 迁移模式 - 支持完整工作空间迁移到新服务器

配置要求

必需配置

GitHub Token:

• 需要一个 Personal Access Token
• 权限要求：repo（仓库读写权限）
• 生成地址：https://github.com/settings/tokens/new

GitHub 仓库:

• 格式：username/repository-name
• 示例：myusername/openclaw-memory-backup
• 建议设为 Private（私有仓库），因为 memory 可能包含敏感信息

配置方式

方案 A：使用环境变量（推荐用于测试）

export GITHUBTOKEN="ghp_xxxxxxxxxxxxxxxxx"
export GITHUB_REPO="yourusername/your-repo"

方案 B：配置文件（推荐用于生产）

在 ~/.openclaw/openclaw.json 中添加：

{
  "skills": {
    "entries": {
      "github-memory-sync": {
        "enabled": true,
        "apiKey": "ghp_xxxxxxxxxxxxxxxxx",
        "env": {
          "GITHUBTOKEN": "ghp_xxxxxxxxxxxxxxxxx",
          "GITHUB_REPO": "username/memory-backup",
          "GITHUB_BRANCH": "main",
          "WORKSPACE_DIR": "/root/.openclaw/workspace"
        }
      }
    }
  }
}

使用示例

首次初始化

用户："初始化 GitHub memory 仓库"
AI: [获取 Token 和仓库信息后执行初始化]

推送更新（备份）

用户："同步到 GitHub" / "备份配置"
AI: [执行推送操作，同步所有记忆和配置文件]

拉取更新（恢复）

用户："从 GitHub 拉取配置" / "恢复备份"
AI: [执行拉取操作，恢复所有文件]

查看状态

用户："检查同步状态"
AI: [显示本地和远程的差异]

🚀 服务器迁移（完整流程）

在原服务器上：

用户："备份所有配置到 GitHub"
AI: [执行完整推送，包括 SOUL.md, IDENTITY.md, USER.md, MEMORY.md, TOOLS.md, memory/* 等]

在新服务器上：

用户："从 GitHub 恢复配置"
AI: [执行以下步骤]
1. 克隆 GitHub 仓库到临时目录
2. 复制所有记忆和配置文件到 workspace
3. 保留新服务器的通道配置（不覆盖 openclaw.json 中的通道凭证）
4. 验证文件完整性

部分恢复

用户："只恢复 MEMORY.md"
AI: [仅拉取指定文件]

用户："恢复 memory 目录"
AI: [仅拉取 memory/*.md 文件]

安全提醒

⚠️ 重要安全注意事项：

1. Token 保护

   • ❌ 不要把 Token 发送到任何公开渠道
   • ❌ 不要在代码中硬编码 Token
   • ✅ 使用环境变量或配置文件
   • ✅ 定期轮换 Token

2. 仓库隐私

   • 🔒 建议将 GitHub 仓库设为 Private（私有）
   • 👁 memory 可能包含敏感信息
   • 📝 审查 memory 内容再上传

3. 权限最小化

   • 只给 Token 必要的权限（repo）
   • 避免使用具有广泛权限的 Token
   • 设置 Token 过期时间（不要永不过期）

激活技能

当用户提到以下关键词时激活此技能：

• "GitHub memory"
• "同步 memory"
• "备份 memory"
• "GitHub 备份"
• "memory 同步"
• "github-memory-sync"

配置流程

1. 获取配置信息

• 向用户询问 GitHub Token
• 向用户询问 GitHub 仓库地址（或帮其创建）

2. 保存配置

• 将 Token 和仓库信息保存到配置文件或环境变量
• 提醒用户注意安全事项

3. 执行操作

• 根据用户请求执行 init/push/pull/status/migrate 操作
• 显示操作结果

4. 验证同步

• 推送后验证远程仓库文件完整性
• 拉取后验证本地文件完整性

注意事项

• 首次使用必须先执行 init 初始化
• 推送前建议先拉取，避免冲突
• 定期检查 Token 是否过期
• 建议启用 GitHub 的两因素认证
• 通道凭证不同步 - openclaw.json 中的 tokens/secrets 不应上传到 GitHub
• 迁移模式 - 使用 migrate 命令可安全恢复到新服务器

相关文件

• scripts/sync_to_github.py - 同步脚本
• references/migration-guide.md - 完整服务器迁移指南
• references/backup-policy.md - 备份策略和安全建议

快速参考

常用命令

# 初始化（首次使用）
python scripts/sync_to_github.py init

# 备份
python scripts/sync_to_github.py push

# 恢复
python scripts/sync_to_github.py pull

# 查看状态
python scripts/sync_to_github.py status

# 迁移到新服务器
BACKUP_DIR=/tmp/openclaw-restore python scripts/sync_to_github.py migrate

对话触发词

• "备份到 GitHub"
• "同步配置"
• "恢复备份"
• "迁移到新服务器"
• "检查同步状态"
• "github-memory-sync"