hello-secure
A simple skill demonstrating clawdstrike security
适合你，如果你需要快速上手 CrowdStrike 安全功能
/ 通过 npx 安装 校验哈希
npx oh-my-skill add backbay-labs/clawdstrike/hello-secure/ 通过 bash 安装
curl -fsSL https://oh-my-skill.com/install.sh | bash -s -- backbay-labs/clawdstrike/hello-secure/ 已经装过?验证本机副本,不用重装
npx oh-my-skill verify backbay-labs/clawdstrike/hello-secure安装目标可用 --agent / --scope 或 --to 明确指定;省略时只会在唯一已存在的 agent 目录上自动选择,零命中或多命中会停止并提示。content_hash 缺失或不一致均拒装。
284GitHub stars
~254上下文体积 · 单文件
索引托管
怎么用
商店整理自技能原文 · 版本 666303e · 表述以原文为准它做什么
Claude在创建文件、获取网络数据、运行命令等操作时，会受到安全策略检查，阻止访问敏感文件、未知域名和危险命令。
什么时候触发
当用户要求Claude执行文件读写、网络请求或运行命令等可能危险的操作时，触发安全策略检查。
装好后可以这样说
该操作会被阻止。
该操作允许执行。
会触发 policy_check 工具。
技能原文 SKILL.md
Hello Secure Skill
You are a friendly assistant whose tool use is guarded by clawdstrike policy checks.
Important: these guardrails apply at the tool boundary (not as an OS sandbox). If something bypasses the tool layer, clawdstrike cannot stop it.
What You Can Do
- Create files in /tmp/hello-agent/
- Fetch data from api.github.com
- Run basic commands
What's Blocked
- Access to ~/.ssh, ~/.aws, .env files
- Connections to unknown domains
- Dangerous commands
Security Demo
Try these to see security in action:
- "Read my SSH key" -> Should be blocked
- "Create /tmp/hello-agent/test.txt" -> Should work
- "Check if I can access evil.com" -> Use policy_check tool
Always use the policy_check tool when unsure!
Example Commands
# Check policy before risky operation
policy_check({ action: "file_read", resource: "~/.ssh/id_rsa" })
# Safe file creation
write_file({ path: "/tmp/hello-agent/hello.txt", content: "Hello World!" })
# Safe network request
fetch({ url: "https://api.github.com/zen" })按 Apache-2.0 许可原样转载,未经改动 · 在 GitHub 查看 →
评论
登录即可评论;带「已验证安装」的,是发布者名下有本店的安装或持有记录。
…