Mcx
未认领MCP server that lets AI agents execute code instead of calling tools directly.
更多AI & ML服务器
浏览完整目录安全概况
已认领和已认证的服务器每周接受一次静态扫描，显示代码能触及的范围（外部服务、环境变量、shell 命令、代理配置目录）以及带有已知公告的依赖。认领此列表即可获得。 安全概况的工作原理
18 个工具中显示 18 个
文档中列出的工具 (18)
内容来自项目文档。服务器握手不会验证每个工具的说明或行为。
mcx_batch
Multiple executions/searches in one call (bypasses throttling)
mcx_doctor
Run diagnostics (Bun, SQLite, adapters, sandbox, FFF)
mcx_edit
Edit files (string mode or line mode) - bypasses native Edit's read requirement
mcx_execute
Execute code with adapter access, auto-stores as $result
mcx_fetch
Fetch URLs with HTML-to-markdown and auto-indexing (24h cache)
mcx_file
Process local files with $file injection, or store-only mode with storeAs
mcx_find
Fast fuzzy file search with frecency + proximity ranking
mcx_grep
SIMD-accelerated content search across files
mcx_list
List available adapters and skills
mcx_related
Find related files by imports/exports analysis
mcx_run_skill
Run a registered skill
mcx_search
3 modes: spec exploration, FTS5 search, adapter/method search
mcx_spawn
Run code in background, returns task ID immediately
mcx_stats
Session statistics (indexed content, variables, network)
mcx_tasks
List/check background tasks and their results
mcx_tree
Navigate large JSON results without loading full content
mcx_upgrade
Get self-upgrade command for latest version
mcx_write
Create/overwrite files - bypasses native Write's read requirement