Token导航 LogoToken导航TokenDH.com
效率敏感数据clawhub未标认证来源可访问clear审计通过

dependency-guard依赖卫士

Agent Skill

dependency-guard 用于补充效率相关能力,适合在 OpenClaw 中需要让 Agent 承接效率相关任务时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

总安装

3,994

周安装

160

GitHub Stars

公开资料未说明

下载量

1,293
OpenClaw

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

MIT-0

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:dependency-guard(依赖卫士)
来源仓库:https://github.com/tuthan/dependency-guard
安装命令:
openclaw skills install dependency-guard
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 OpenClaw 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

ClawHubOpenClaw
openclaw skills install dependency-guard

简介

dependency-guard 在添加、升级或删除依赖时应用供应链安全护栏。

  • 适合在变更前自动阻断高风险包或触发人工审核流程。
  • 通过 clawhub 安装并使用 openclaw skills install dependency-guard 命令部署。
  • 需配置信任源白名单和许可证黑名单规则。
  • 建议与 CI 系统集成,实现提交即检的防护机制。

SKILL.md

name
dependency-guard
version
1.0.1
description
Use when a task adds, upgrades, removes, or reviews software dependencies and the agent should apply a Socket-based supply-chain guardrail before changing manifests or lockfiles. Prefer MCP depscore when available, otherwise use the bundled Socket CLI helper. Stop and recommend an alternative or human review when risk signals are weak.
metadata
{"openclaw":{"emoji":"🛡️","requires":{"bins":["socket"]}}}

Dependency Guard

Use this skill when dependency changes are in scope for npm, pnpm, yarn, Python packages, or other package ecosystems supported by Socket.

Prerequisites

  • The socket CLI must be installed and on PATH (npm install -g socket).
  • Authentication is required for CLI-based reviews. See the Authentication section below.

Workflow

  1. Confirm the exact dependency change being proposed.
  2. Check whether the feature can be implemented with the standard library or an existing project dependency.
  3. Prefer MCP depscore if the host agent exposes it.
  4. Otherwise run scripts/check_dependency.sh <ecosystem> <package> [version].
  5. Apply the policy in references/policy.md.
  6. Apply the decision rules in references/decision-matrix.md.
  7. Before making the change, report:

- why the package is needed - whether an existing alternative exists - what Socket reported - whether install scripts, risky capabilities, or transitive risk are present

  1. If the decision is allow_with_warning, present the warning clearly before making the change. If the decision is block_pending_human_review or block, stop and propose either:

- a safer dependency - a no-dependency implementation - explicit human review

Authentication

Three authentication paths are supported, in order of preference:

  1. MCP depscore — no local credentials needed; works through the host agent's MCP connection.
  2. socket login — interactive CLI login; stores auth locally.

- If your CLI supports it, pressing Enter at the token prompt uses limited public access. - To use a private token, paste it at the prompt instead.

  1. SOCKET_SECURITY_API_TOKEN env var — set this for CI or headless environments.
Security: Never paste private tokens into agent prompts. Use the env var or socket login instead.
CI note: GitHub Actions workflows use SOCKET_SECURITY_API_KEY (a separate GitHub-integration key), not SOCKET_SECURITY_API_TOKEN. See examples/github/dependency-guard.yml.

Reporting Contract

Use the short response template in references/examples.md when presenting the package review to the user.

References

  • Read references/policy.md for the canonical guardrail.
  • Read references/decision-matrix.md for allow/block criteria.
  • Read references/examples.md for user-facing review examples.

Notes

  • Keep SKILL.md lean; do not duplicate the full policy here.
  • OpenClaw and ClawHub expect metadata to be a single-line JSON object in frontmatter, so keep the OpenClaw metadata compact.
  • The version field in frontmatter is the single source of truth; use publish_clawhub.sh --bump patch|minor|major to auto-increment.
  • Do not assume system-wide wrapper enforcement or shell-completion setup is desirable; keep CLI setup minimal.
  • If Socket tooling is unavailable, require human review before adding the dependency.
  • Review manifest and lockfile changes together.

适合场景

01

OpenClaw 用户查找和安装 Skill 时

02

用户想查找某类 Agent Skill 时

03

需要根据任务场景推荐可安装能力包时

04

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

补充不同宿主或平台的使用分布数据

能力 5

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

OpenClaw

74.03%
按下载量换算957

安全审计

VirusTotal

未展示

ClawScan

通过

Static analysis

通过

权限和风险

敏感数据

该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills