← 返回 Skills 市场
338
总下载
0
收藏
0
当前安装
2
版本数
在 OpenClaw 中安装
/install env-secure-manager
功能描述
Secure environment variable & secret management with AES-256 encryption, auto-redaction, permission control, prevent credential leakage
使用说明 (SKILL.md)
🔐 环境变量安全管理器
核心亮点
- 🛡️ AES-256加密存储：敏感信息自动加密，即使配置文件泄露也无法获取明文密钥
- 🚫 自动脱敏机制：自动检测并脱敏输出中的敏感信息，防止密钥泄露到日志/聊天记录
- 🔑 权限控制：访问敏感值需要显式授权，避免意外泄露
- 🔄 自动密钥生成：首次使用自动生成安全密钥，也支持自定义密钥
🎯 适用场景
- 管理API密钥、数据库密码等敏感信息
- 防止敏感信息泄露到日志、输出或会话历史
- 批量加载环境变量，统一管理配置
- 多Agent环境下的安全配置共享
📝 参数说明
| 参数 | 类型 | 必填 | 说明 |
|---|---|---|---|
| action | string | 是 | 操作类型：init/set/get/list/delete/redact/loadFromEnv |
| key | string | 否 | set/get/delete操作必填，环境变量名，大写字母+下划线 |
| value | string | 否 | set操作必填，变量值 |
| isSecret | boolean | 否 | set操作可选，是否为敏感信息，默认false |
| allowSecret | boolean | 否 | get操作可选，是否允许获取敏感值，默认false |
| text | string | 否 | redact操作必填，要脱敏的文本 |
| prefix | string | 否 | loadFromEnv操作可选，环境变量前缀，默认OPENCLAW_ |
| encryptionKey | string | 否 | init操作可选，自定义32位加密密钥 |
💡 开箱即用示例
初始化（可选，建议启动时调用）
// 使用自定义密钥
await skills.envSecureManager({
action: "init",
encryptionKey: "你的32位安全密钥"
});
// 自动生成密钥
await skills.envSecureManager({ action: "init" });
存储敏感密钥
await skills.envSecureManager({
action: "set",
key: "OPENAI_API_KEY",
value: "sk-xxx",
isSecret: true
});
安全获取密钥
const result = await skills.envSecureManager({
action: "get",
key: "OPENAI_API_KEY",
allowSecret: true // 必须显式授权才能获取敏感值
});
输出自动脱敏
// 即使日志里不小心打印了密钥，也会自动脱敏
const logText = `调用OpenAI API，密钥是sk-xxx，参数是xxx`;
const redacted = await skills.envSecureManager({
action: "redact",
text: logText
});
// 输出：调用OpenAI API，密钥是***REDACTED***，参数是xxx
🔧 技术实现说明
- 使用AES-GCM 256位加密算法，符合企业级安全标准
- 敏感信息永远不以明文存储，运行时解密
- 自动脱敏机制支持多值替换，覆盖所有泄露场景
- 轻量无依赖，不影响Agent执行性能
安全使用建议
This skill is plausibly what it claims to be, but there are a few red flags and bugs you should consider before installing: (1) the code reads and writes process environment variables (OPENCLAW_ENV_ENCRYPTION_KEY and loadFromEnv reads the environment). If you don't want the skill to access your environment, do not enable it. (2) The module fetches dependencies from deno.land at runtime — review those remote packages or run in an environment with restricted network access. (3) There is a likely implementation bug: loadFromEnv spreads the encrypt() result but doesn't set the stored 'value' field correctly, which can cause runtime errors or unexpected behavior. (4) Prefer providing an explicit encryptionKey when initializing rather than relying on auto-generated keys that the skill writes into the process env. (5) Test this in an isolated workspace (or with non-production secrets) and review/modify the source if you need stricter guarantees (for example, restrict env access or remove Deno.env.set). If you want me to, I can point to the exact lines with the bug and propose a corrected code snippet.
功能分析
Type: OpenClaw Skill
Name: env-secure-manager
Version: 1.0.1
The env-secure-manager skill provides legitimate utility for managing sensitive environment variables using AES-256-GCM encryption and automated redaction. The implementation in index.ts uses standard Deno crypto libraries and follows the documented behavior without any evidence of data exfiltration, unauthorized access, or malicious intent.
能力评估
Purpose & Capability
Name/description (AES-256 secret storage, redaction, permissions) align with the included code: the module encrypts/decrypts values, redacts secrets from text, and can load environment variables. However SKILL.md claims 'lightweight no dependencies' yet the code imports zod and std modules from deno.land; also the metadata declares no required env vars but the code reads/writes an OPENCLAW_ENV_ENCRYPTION_KEY environment variable.
Instruction Scope
The SKILL.md documents loadFromEnv, set/get/redact actions, which matches code. But the implementation reads Deno.env.toObject() (iterates process environment) and will set Deno.env.set('OPENCLAW_ENV_ENCRYPTION_KEY', ...) when auto-generating a key. The metadata did not declare that the skill will read or write environment variables. Reading all env entries (even though it filters by prefix) and writing a process env variable are broader scopes than the registry metadata indicates.
Install Mechanism
No install spec in registry, but the code includes remote imports from deno.land (zod and std modules). Fetching runtime dependencies from deno.land is common for Deno but it means remote code will be downloaded/executed at runtime — moderate supply-chain risk compared to fully local code.
Credentials
The skill declares no required env vars, yet init() will read OPENCLAW_ENV_ENCRYPTION_KEY and may write it to the process environment. loadFromEnv iterates the entire environment (via Deno.env.toObject()) and will import keys with a given prefix. This behavior is proportionate to a secret manager only if callers expect the skill to access process env; but that access is not declared and could expose many environment variables if the prefix is broad or changed.
Persistence & Privilege
The skill does not request 'always:true' and does not modify other skills. It does call Deno.env.set to persist the auto-generated encryption key into the process environment, which changes runtime state and could influence other components. This is not necessarily malicious but is a persistence/side-effect the user should be aware of.
如何使用
- 确保已安装 OpenClaw（本地或 Docker 部署）
- 在对话框中输入安装命令：
/install env-secure-manager - 安装完成后，直接呼叫该 Skill 的名称或使用
/env-secure-manager触发 - 根据 Skill 的参数说明提供必要输入，即可获得结构化输出
版本历史
v1.0.1
- Removed the skill.yaml file.
- Metadata previously in skill.yaml is now moved into SKILL.md for a simplified file structure.
- No changes to features or core documentation content.
v1.0.0
Initial release of env-secure-manager.
- Securely manages environment variables and secrets with AES-256 encryption.
- Auto-redacts sensitive information from outputs to prevent credential leakage.
- Provides permission control: explicit authorization required to access sensitive values.
- Automatically generates encryption keys on first use, with support for custom keys.
- Suitable for managing API keys, passwords, and environment configs in multi-agent setups.
元数据
常见问题
env-secure-manager 是什么？
Secure environment variable & secret management with AES-256 encryption, auto-redaction, permission control, prevent credential leakage. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 338 次。
如何安装 env-secure-manager？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install env-secure-manager」即可一键安装，无需额外配置。
env-secure-manager 是免费的吗？
是的，env-secure-manager 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
env-secure-manager 支持哪些平台？
env-secure-manager 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（cross-platform）。
谁开发了 env-secure-manager？
由 Ayalili（@ayalili）开发并维护，当前版本 v1.0.1。
推荐 Skills