KillerSkills
Security verdicts for agent skills and MCP servers, from sandbox runs, before you install them.
作者提供的说明(英文原文)
Scan, monitor and fix a live web app from your editor: graded security reports with fix prompts.
这三项永远不会合并成一个分数:意见、测量和人气是不同的东西。
我们还没有为它写通俗的介绍。下面是我们从注册中心读取到的事实。
选择你使用的 AI 应用,然后复制设置。
Authorization标为机密的值是你账户的钥匙。千万不要分享给别人,也不要粘贴到聊天里。
你需要
你用的是哪个应用?
在上方选择你的应用。我们会记住它,下次按钮会显示“安装到 …”。
这个地址需要在请求的 header 中放入密钥,而添加连接器的菜单无法设置。请改用 Claude Code、Cursor 或 VS Code。
在终端中运行这条命令
claude mcp add --transport http secscan https://secscan.us/api/mcp --header 'Authorization: Bearer {token}'运行前请先填写空着的值。
粘贴到项目中的 .cursor/mcp.json,或 ~/.cursor/mcp.json
{
"mcpServers": {
"secscan": {
"url": "https://secscan.us/api/mcp",
"headers": {
"Authorization": "Bearer {token}"
}
}
}
}浏览器会先询问是否打开应用,然后应用会显示设置并请你确认。密钥的值会留空,由你自己填写。
粘贴到项目中的 .vscode/mcp.json
{
"servers": {
"secscan": {
"type": "http",
"url": "https://secscan.us/api/mcp",
"headers": {
"Authorization": "Bearer {token}"
}
}
}
}浏览器会先询问是否打开应用,然后应用会显示设置并请你确认。密钥的值会留空,由你自己填写。
这个地址需要在请求的 header 中放入密钥,而添加连接器的菜单无法设置。请改用 Claude Code、Cursor 或 VS Code。
你的选择通过 Cookie 保存在这台设备上,无需账户。
Authorization 机密 必填 HTTP header这些设置是根据作者在注册中心的条目生成的。如果 README 的写法不同,请以 README 为准。
工具名称和描述与服务器在 2026年11月9日 报告的完全一致。这些内容由作者撰写,不是我们写的;我们以纯文本形式显示。
scan_urlStart a SecScan security scan of a web application the user owns or is authorised to test. Returns a scan_id; most scans finish in under a minute — then call get_scan_status with wait_seconds, or get_report. Active tests (injection, XSS, SSRF…) run only on domains the user has ve
get_scan_statusStatus of a scan (queued, scanning, analyzing, complete, failed). With wait_seconds (max 60) it waits for the scan to finish and returns the full report as soon as it does.
get_reportThe finished report for a scan: grade, what the scan tested and what it skipped (a clean grade says nothing about skipped areas, so say so), prioritised findings with fixes and evidence, and a fix prompt written for the user's AI editor. Findings come 25 per page, most severe fir
list_recent_scansThe user's most recent scans with their status, newest first.
list_verified_domainsDomains the user has proved they own. Only these receive active testing (injection, XSS, SSRF, access control); others get passive checks. Verify more at https://secscan.us/domains.
get_accountHow many scans the user can still run — free scans, plan scans and credits — and their plan. Check this before starting several scans.
start_domain_verificationBegin proving the user owns a domain, which unlocks active tests (injection, XSS, SSRF, access control) on its scans. Returns a file to publish on the site, or a DNS TXT record — an editor can usually add the file to the codebase and deploy it. Then call check_domain_verification
check_domain_verificationCheck whether the file or DNS record from start_domain_verification is live. On success the domain is verified and its next scan includes active tests. DNS changes can take a few minutes.
list_monitorsSites under continuous monitoring: latest grade, last and next scan, uptime check, CVE alerts, new problems in the last scan and certificate expiry. Use the monitor id with monitor_scan_now.
add_monitorPut a site the user owns under continuous monitoring: hourly uptime checks, CVE matching, certificate alerts and regular rescans. Runs a full baseline scan straight away, which uses one of the user's scans exactly as in the app (free for plan holders). Returns the baseline scan_i
monitor_scan_nowRun a full scan of a monitored site now instead of waiting for its schedule — e.g. to confirm a fix. Free for plan holders; otherwise uses one of the user's scans, as in the app. Takes the monitor id from list_monitors.
dismiss_findingMark a finding as a false positive for this site, so future scans of it stop reporting it — the same as Dismiss in the app, and undoable there. ONLY use this after the user has confirmed the finding is wrong; never dismiss a real problem to improve a grade.
我们检查的是任何人都能看到的信息;我们没有阅读代码。不适用的检查项不计入总数。 · 检查于 2026年11月9日
注册中心没有列出源代码仓库,因此无法将发布者(secscan.us)与代码对应起来
我们还没有从 GitHub 读取这个仓库
我们还没有从 GitHub 读取这个仓库
在 282 毫秒内响应(HTTP 200)
无需登录
已读取 12 个工具描述,未发现针对 AI 的隐藏指令
与注册中心里这个分类中所有活跃的服务器比较。只有当发布者拥有该仓库时,星标数才计入。
它在 3 项适用的基础检查中通过了 2 项。我们没有阅读代码。在添加密钥或令牌之前,请先看看这个服务器要求哪些权限,并使用只具备所需权限的密钥。
你需要填写:Authorization
它在作者的云端运行。通过互联网连接,无需安装,但你发送的内容会经过作者的服务器。
GitHub 仓库最后更新于–,最新版本(1.1.0)于2个月前进入注册中心。
支持 MCP 的应用,例如 Claude Desktop、Claude Code、Cursor 和 VS Code。它通过 URL 连接,所以接受远程 MCP 服务器的应用也可以使用,例如开启开发者模式的 ChatGPT。
同一分类中有:KillerSkills, OpenWorkProof, Shipshape MCP
同一分类中其他作者的服务器,按星标数排序。
如何看卡片:↓ 每周下载量i★ GitHub 星标i三大支柱:评价 · 检查 · 使用情况i更新 = 最新版本发布到注册中心的时间i
Security verdicts for agent skills and MCP servers, from sandbox runs, before you install them.
Verifiable agent evidence; opt-in Codex guard and reference SOC alert review.
Dockerfile and Kubernetes lint for AI agents: Pod Security Standards, A–F grade, a fix per finding.
仓库中没有找到 README