Patterns and techniques for evaluating and improving AI agent outputs.
日本語の概要は準備中です。原文の説明を表示しています。
Validates MCP server connectivity and health. Checks stdio and http MCP servers for availability, reports connection status, and provides fallback guidance. Triggers on: mcp health, check mcp, validate mcp, mcp status, mcp connectivity.
インストールする前に、エージェントに与えられる指示の中身を確認できます。
Validates MCP server connectivity and health status for local stdio and remote http MCP servers.
Always load before analysis:
# Check all MCP servers
[mcp-health:check-all]
# Check specific MCP
[mcp-health:check microsoftdocs-mcp]
# Get health report
[mcp-health:report]
Launch local processes via stdio transport:
Connect to remote APIs via HTTP:
# Check if command exists and is executable
which uvx || which npx
# Test stdio connection
echo '{"jsonrpc":"2.0","id":1,"method":"initialize","params":{}}' | \
uvx --from serena-mcp server 2>&1 | head -5
# Test HTTP connectivity
curl -s -o /dev/null -w "%{http_code}" https://learn.microsoft.com/api/mcp/health
curl -s -o /dev/null -w "%{http_code}" https://mcp.deepwiki.com/health
Add to session start hooks:
# Inline mcp-health-check.sh: Validate MCP servers
set -uo pipefail
MCP_CONFIG="${MCP_CONFIG:-.rulesync/mcp.json}"
[ ! -f "$MCP_CONFIG" ] && exit 0
# Check if jq is available
if ! command -v jq >/dev/null 2>&1; then
echo "{\"warning\":\"jq not found - skipping MCP health check\"}"
exit 0
fi
# Extract and validate MCP servers
FAILED_MCP=""
while IFS= read -r mcp_name; do
mcp_type=$(jq -r ".mcpServers.\"$mcp_name\".type" "$MCP_CONFIG" 2>/dev/null || echo "")
case "$mcp_type" in
http)
url=$(jq -r ".mcpServers.\"$mcp_name\".url" "$MCP_CONFIG" 2>/dev/null)
if [ -n "$url" ]; then
if ! curl -s --connect-timeout 5 -o /dev/null "$url/health" 2>/dev/null; then
FAILED_MCP="$FAILED_MCP $mcp_name"
fi
fi
;;
stdio)
# Check if command exists
cmd=$(jq -r ".mcpServers.\"$mcp_name\".command" "$MCP_CONFIG" 2>/dev/null)
if ! command -v "$cmd" >/dev/null 2>&1; then
# Check common package managers
case "$cmd" in
uvx) command -v uvx || command -v pipx || FAILED_MCP="$FAILED_MCP $mcp_name" ;;
npx) command -v npx || FAILED_MCP="$FAILED_MCP $mcp_name" ;;
esac
fi
;;
esac
done < <(jq -r '.mcpServers | keys[]' "$MCP_CONFIG" 2>/dev/null || true)
if [ -n "$FAILED_MCP" ]; then
echo "{\"warning\":\"MCP servers unavailable:$FAILED_MCP\"}"
else
echo "{\"status\":\"healthy\"}"
fi
exit 0
{
"timestamp": "2026-02-28T10:45:00Z",
"servers": {
"serena": {
"type": "stdio",
"status": "healthy",
"command": "uvx",
"available": true
},
"microsoftdocs-mcp": {
"type": "http",
"status": "healthy",
"url": "https://learn.microsoft.com/api/mcp",
"latency_ms": 45
},
"context7": {
"type": "http",
"status": "healthy",
"url": "https://mcp.context7.com/mcp"
}
},
"summary": {
"total": 5,
"healthy": 4,
"unavailable": 1
}
}
Problem: Command not found (uvx, npx)
Solutions:
# Install uvx via pipx
pipx install uvx
# Install npx (comes with Node.js)
npm install -g npx
# Or use npm directly for npx packages
npm install -g @anthropic-ai/mcp-server-fetch
Problem: MCP server crashes immediately
Solutions:
--verbose flagpipx upgrade serena-mcpProblem: Connection timeout
Solutions:
curl -I <mcp-url>Problem: 401 Unauthorized
Solutions:
.env or environmentWhen MCP servers are unavailable:
Add to .rulesync/hooks.json:
{
"hooks": {
"sessionStart": [
{
"type": "command",
"command": "# MCP health check script here",
"timeout": 30
}
]
}
}
Reference in skills that depend on MCP:
## Preloaded Skills
- [skill:mcp-health] -- validate MCP availability
- [skill:microsoft-learn-mcp] -- query Microsoft documentation
まだレビューはありません。使ってみた感想をお寄せください。
概要と使いどころ
Patterns and techniques for evaluating and improving AI agent outputs.
日本語の概要は準備中です。原文の説明を表示しています。
Comprehensive AI prompt engineering safety review and improvement prompt. Analyzes prompts for safety, bias, security vulnerabilities, and effectiveness while providing detailed improvement recommendations.
日本語の概要は準備中です。原文の説明を表示しています。
Use when user requests research requiring multiple sources, comprehensive analysis, or synthesis across topics - technical research, domain knowledge gathering, market analysis, or learning about complex subjects
日本語の概要は準備中です。原文の説明を表示しています。
AI-powered wiki generation for code repositories with commands, agents, and skills
日本語の概要は準備中です。原文の説明を表示しています。
Use when building .NET 10 or C# 14 applications; when using minimal APIs, modular monolith patterns, or feature folders; when implementing HTTP resilience, Options pattern, Channels, or validation; when seeing outdated patterns like old extension method syntax
日本語の概要は準備中です。原文の説明を表示しています。
Implements accessible .NET UI. SemanticProperties, ARIA, AutomationPeer, testing per platform.
日本語の概要は準備中です。原文の説明を表示しています。