MCP server that monitors system health — CPU, memory, disk, Chrome/Node processes, Docker, and network — via shell commands.
12 tools with basic schemas and descriptions, but significant gaps in parameter documentation and output schema clarity. Most tools lack descriptions for their parameters (e.g., 'autoFix' in health_check_all has a description, but many boolean flags lack context). No documented output schemas visible, LLMs cannot predict what fields to expect. Several tools combine multiple concerns (health_check_all runs comprehensive checks; cleanup_resources targets multiple cleanup types). Naming is generally verb-first and clear, but some tools like 'monitor_start' lack clarity on what 'background monitoring' entails. Error handling and recovery guidance absent from descriptions.
Check and manage Chrome/Chromium processes
Check Docker containers status
Check health of all MCP servers
Check network connectivity to key services
Check Node.js processes (MCP servers, etc.)
Check CPU, memory, and disk usage
Clean up system resources (temp files, caches, etc.)
No output schemas documented. LLMs cannot predict return structure (fields, types, pagination). Tools like check_system_resources and check_mcp_servers lack any indication of what data they return.
Multiple tools combine unrelated concerns. health_check_all runs 'comprehensive health check on all systems', unclear what 'all' means. cleanup_resources targets 4 different cleanup types (temp, npm-cache, chrome-cache, logs) in one tool. Should split into focused, single-responsibility tools.
Destructive operations (kill_process, cleanup_resources) lack confirmation/dry-run guidance in descriptions. No mention of how to safely preview changes before execution. LLMs may execute irreversible actions without understanding consequences.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 51 | 2026-07-28+ | v2 |
Get health check history from log
Get process tree for Claude-related processes
Run comprehensive health check on all systems
Kill a process by PID or pattern
Start background monitoring (writes to log file)
Parameter descriptions are minimal or absent. 'showDetails' in check_node_processes says 'Show detailed process info' but doesn't explain what 'detailed' means or what extra fields appear. 'interval' in monitor_start lacks guidance on valid range (1-3600? 1-86400?).
No error handling or recovery guidance in tool descriptions. If kill_process fails (PID not found, permission denied), LLM has no guidance on what to try next. No mention of retryability or error categories.