← 返回 Skills 市场
rohitg00

Kubernetes Skills

作者 Rohit Ghumare · GitHub ↗ · v1.0.0
cross-platform ⚠ suspicious
2592
总下载
0
收藏
4
当前安装
1
版本数
在 OpenClaw 中安装
/install k8s-browser
功能描述
Browser automation for Kubernetes dashboards and web UIs. Use when interacting with Kubernetes Dashboard, Grafana, ArgoCD UI, or other web interfaces. Requires MCP_BROWSER_ENABLED=true.
使用说明 (SKILL.md)

Browser Automation for Kubernetes

Automate Kubernetes web UIs using kubectl-mcp-server's browser tools (26 tools).

Enable Browser Tools

export MCP_BROWSER_ENABLED=true

# Optional: Cloud provider
export MCP_BROWSER_PROVIDER=browserbase  # or browseruse
export BROWSERBASE_API_KEY=bb_...

Basic Navigation

# Open URL
browser_open(url="http://localhost:8001/api/v1/namespaces/kubernetes-dashboard/services/https:kubernetes-dashboard:/proxy/")

# Open with auth headers
browser_open_with_headers(
    url="https://grafana.example.com",
    headers={"Authorization": "Bearer token123"}
)

# Navigate
browser_navigate(url="https://argocd.example.com/applications")

# Go back/forward
browser_back()
browser_forward()

# Refresh
browser_refresh()

Screenshots and Content

# Take screenshot
browser_screenshot(path="dashboard.png")

# Full page screenshot
browser_screenshot(path="full-page.png", full_page=True)

# Get page content
browser_content()

# Get page title
browser_title()

# Get current URL
browser_url()

Interactions

# Click element
browser_click(selector="button.submit")
browser_click(selector="text=Deploy")
browser_click(selector="#sync-button")

# Type text
browser_type(selector="input[name=search]", text="my-deployment")
browser_type(selector=".search-box", text="nginx")

# Fill form
browser_fill(selector="#namespace", text="production")

# Select dropdown
browser_select(selector="select#cluster", value="prod-cluster")

# Press key
browser_press(key="Enter")
browser_press(key="Escape")

Waiting

# Wait for element
browser_wait_for_selector(selector=".loading", state="hidden")
browser_wait_for_selector(selector=".data-table", state="visible")

# Wait for navigation
browser_wait_for_navigation()

# Wait for network idle
browser_wait_for_load_state(state="networkidle")

Session Management

# List sessions
browser_session_list()

# Switch session
browser_session_switch(session_id="my-session")

# Close browser
browser_close()

Viewport and Device

# Set viewport size
browser_set_viewport(width=1920, height=1080)

# Emulate device
browser_set_viewport(device="iPhone 12")

Kubernetes Dashboard Workflow

# 1. Start kubectl proxy
# kubectl proxy &

# 2. Open dashboard
browser_open(url="http://localhost:8001/api/v1/namespaces/kubernetes-dashboard/services/https:kubernetes-dashboard:/proxy/")

# 3. Navigate to workloads
browser_click(selector="text=Workloads")

# 4. Take screenshot
browser_screenshot(path="workloads.png")

# 5. Search for deployment
browser_type(selector="input[placeholder*=search]", text="nginx")
browser_press(key="Enter")

Grafana Dashboard Workflow

# 1. Open Grafana
browser_open_with_headers(
    url="https://grafana.example.com/d/k8s-cluster",
    headers={"Authorization": "Bearer admin-token"}
)

# 2. Set time range
browser_click(selector="button[aria-label='Time picker']")
browser_click(selector="text=Last 1 hour")

# 3. Screenshot dashboard
browser_screenshot(path="grafana-cluster.png", full_page=True)

ArgoCD UI Workflow

# 1. Open ArgoCD
browser_open(url="https://argocd.example.com")

# 2. Login
browser_fill(selector="input[name=username]", text="admin")
browser_fill(selector="input[name=password]", text="password")
browser_click(selector="button[type=submit]")

# 3. Navigate to app
browser_wait_for_selector(selector=".applications-list")
browser_click(selector="text=my-application")

# 4. Sync application
browser_click(selector="button.sync-button")
browser_click(selector="text=Synchronize")

Related Skills

安全使用建议
This skill’s behavior (opening arbitrary dashboard URLs, attaching auth headers, taking screenshots, and retrieving page content) is consistent with a browser automation tool, but the SKILL.md references environment variables and external provider API keys that are not declared in the registry metadata and the package has no known source/homepage. Before installing, verify the author/source and ask for: an explicit list of required env vars, where captured screenshots/page content are stored or transmitted, network endpoints contacted (especially for 'browserbase'/'browseruse'), and the exact data retention/usage policy. Avoid supplying real cluster/admin credentials or bearer tokens until you confirm the skill keeps data local or you trust the external provider. If you must test, do so in an isolated environment with non-production credentials and monitor network traffic. Consider disabling autonomous invocation until you can confirm there is no exfiltration to external services.
功能分析
Type: OpenClaw Skill Name: k8s-browser Version: 1.0.0 The skill bundle provides extensive browser automation capabilities, including opening arbitrary URLs, injecting custom headers (e.g., Authorization tokens), filling forms with credentials, retrieving full page content, and taking screenshots. While these capabilities are plausibly needed for its stated purpose of 'Browser automation for Kubernetes dashboards and web UIs', they represent significant security risks. The `SKILL.md` file explicitly demonstrates the use of sensitive data like `admin-token` and `password` in examples, highlighting the potential for an AI agent to be prompted to access or manipulate sensitive internal web interfaces, leading to unauthorized data exposure or actions if misused via prompt injection.
能力评估
Purpose & Capability
The skill claims to be a browser automation helper for Kubernetes dashboards, which aligns with the instructions. However the description and SKILL.md say MCP_BROWSER_ENABLED must be set and reference optional provider credentials (MCP_BROWSER_PROVIDER, BROWSERBASE_API_KEY), but the registry metadata lists no required environment variables. That mismatch (declared 'none' vs. instructions requiring env flags/keys) is incoherent and unexplained. Also the skill has no source/homepage, reducing transparency.
Instruction Scope
SKILL.md contains explicit runtime commands for opening arbitrary URLs, attaching Authorization headers, taking screenshots, and retrieving page content. Those are within the stated purpose, but they allow collection and export of potentially sensitive cluster/dashboard data and credentials (the examples even show filling login forms and using bearer tokens). The instructions do not constrain network destinations or explain whether captured screenshots/content remain local or are sent to an external provider, which is a scope and data-exfiltration risk.
Install Mechanism
This is an instruction-only skill with no install spec and no code files — minimal disk footprint. That lowers risk compared to arbitrary downloads. However SKILL.md references external browser provider services (browserbase/browseruse) without an install or network-policy description, so it's unclear how those services are contacted or whether they will receive captured data.
Credentials
The SKILL.md asks users to export MCP_BROWSER_ENABLED and optionally MCP_BROWSER_PROVIDER and BROWSERBASE_API_KEY, but the registry lists no required env vars or primary credential. The skill also shows usage of Authorization headers and credentials in examples without declaring or justifying where such secrets should come from or how they are protected. Requesting or using external provider API keys and arbitrary auth headers is disproportionate without clear justification or constraints.
Persistence & Privilege
always is false and there is no install step that modifies agent/system configuration. The skill does not request elevated or permanent presence. Autonomous invocation is allowed (platform default) — combine this with the other concerns when deciding whether to enable auto-invoke.
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install k8s-browser
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /k8s-browser 触发
  4. 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
v1.0.0
- Initial release of k8s-browser skill for automating Kubernetes dashboards and web UIs. - Provides 26 browser automation tools for navigating, interacting, and capturing screenshots of Kubernetes Dashboard, Grafana, ArgoCD UI, and similar interfaces. - Requires enabling browser features with MCP_BROWSER_ENABLED=true. - Includes example workflows for Kubernetes Dashboard, Grafana, and ArgoCD UI automation. - Supports session management, device emulation, and advanced selectors for clicks and form fills.
元数据
Slug k8s-browser
版本 1.0.0
许可证
累计安装 4
当前安装数 4
历史版本数 1
常见问题

Kubernetes Skills 是什么?

Browser automation for Kubernetes dashboards and web UIs. Use when interacting with Kubernetes Dashboard, Grafana, ArgoCD UI, or other web interfaces. Requires MCP_BROWSER_ENABLED=true. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 2592 次。

如何安装 Kubernetes Skills?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install k8s-browser」即可一键安装,无需额外配置。

Kubernetes Skills 是免费的吗?

是的,Kubernetes Skills 完全免费(开源免费),可自由下载、安装和使用。

Kubernetes Skills 支持哪些平台?

Kubernetes Skills 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 Kubernetes Skills?

由 Rohit Ghumare(@rohitg00)开发并维护,当前版本 v1.0.0。

💬 留言讨论