Enables instant approval notifications and links for high-risk operations through Telegram bot integration, allowing human approval via Telegram.
This server does not currently have a deployable release on Glama. Follow the repository's setup instructions to install and configure it for your MCP client, including any required credentials and local dependencies.
MCP Guard Server — Budget control, approval workflows, and audit logging for AI agents.
Works with Claude Code, Cursor, ChatGPT, and any AI agent that speaks the MCP protocol.
Budget Control — Daily credits hard limit prevents agent overspending
High-Risk Approval — Operations exceeding amount threshold or matching sensitive resource keywords trigger human approval
Approval Workflow — Agent requests → token generated → human approves via Dashboard or Telegram
Audit Log — All spend, approval, and rejection actions recorded with cursor-based pagination
Web Dashboard — Gin + HTMX + Tailwind real-time dashboard
Telegram Notifications — Instant approval links for high-risk operations
Multi-Transport — stdio / SSE / StreamableHTTP with automatic detection
Related MCP server: llm-cost-guard
Tool | Description |
| Query today's budget status |
| Execute a spend (auto-detects high-risk) |
| Explicitly request human approval |
| Approve a pending token |
| Reject a pending token |
| Poll approval status |
| Query audit log with filtering and pagination |
| List all pending approval requests |
Go 1.24+
No CGO required (pure Go SQLite driver)
git clone https://github.com/dygogogo/agent-guard-mcp.git
cd agent-guard-mcp
go build -o mcp-guard main.goDownload the latest release for your platform:
Platform | amd64 | arm64 |
macOS | ||
Linux | ||
Windows |
Or visit the latest release page.
Configure via environment variables or .env file:
Variable | Description | Default |
| Transport: stdio / sse / http | Auto-detect (TTY → http) |
| Daily budget cap (credits) | 10.0 |
| High-risk amount threshold | 2.0 |
| High-risk resource keywords (comma-separated) | delete,send |
| SQLite database path | ./mcp-guard.db |
| Dashboard HTTP port | 8080 |
| Base URL for approval links | |
| Log level: debug/info/warn/error | info |
| Payer identity | hostname |
| Telegram Bot Token (optional) | - |
| Telegram Chat ID (optional) | - |
# HTTP mode (auto-detected, with Dashboard)
./mcp-guard
# Dashboard: http://localhost:8080/dashboard
# MCP endpoint: http://localhost:8080/mcp
# stdio mode (for MCP clients)
MCP_TRANSPORT=stdio ./mcp-guard
# SSE mode
MCP_TRANSPORT=sse ./mcp-guard
# SSE endpoint: http://localhost:8080/sseAdd to Claude Code's MCP configuration:
{
"mcpServers": {
"agent-guard-mcp": {
"command": "mcp-guard",
"env": {
"MCP_TRANSPORT": "stdio",
"BUDGET_LIMIT": "10"
}
}
}
}┌─────────────┐ MCP Protocol ┌────────────────┐
│ AI Agent │ ◄──────────────────► │ MCP Guard │
│ (Claude, │ stdio / HTTP │ Server │
│ Cursor...) │ │ │
└─────────────┘ │ ┌────────────┐ │
│ │ BudgetStore │ │
┌─────────────┐ HTTP │ │ (SQLite) │ │
│ Dashboard │ ◄──────────────────► │ └────────────┘ │
│ (Gin+HTMX) │ └────────────────┘
└─────────────┘
┌─────────────┐ Webhook
│ Telegram │ ◄────── Approval notifications
└─────────────┘File | Description |
| Entry point, transport selection, graceful shutdown |
| MCP Server with 8 registered tools |
| BudgetStore interface + SQLite implementation |
| High-risk detection, approval workflow, Telegram |
| Environment config, auto transport detection |
| zap logging (stdio mode: file only) |
| Gin Web Dashboard |
1. Agent calls spend(amount=5.0, resource="/api/delete")
2. MCP Guard detects high-risk (amount > threshold OR resource keyword matched)
3. Returns {status: "pending_approval", token: "xxx"}
4. Agent polls check_approval(token) for status
5. Human approves/rejects via Dashboard or Telegram
6. Agent receives final result (approved/rejected/budget_exceeded)# All tests with race detection
go test -race -count=1 ./...
# Integration tests only
go test -race -run TestIntegration -v ./...
# Coverage
go test -race -cover ./...Go 1.24 — Language
mcp-go — MCP protocol Go SDK
Gin — Web framework (Dashboard)
modernc.org/sqlite — Pure Go SQLite (no CGO)
zap — Structured logging
HTMX + Tailwind CSS — Dashboard frontend
MIT
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
This server cannot be deployed
AgentGuard — 20-tool AI safety MCP: policy preflight, risk scoring, audit logging, rate limits.
Zero-secret MCP gateway for AI agents: risk-scored, audited calls with human-in-the-loop approval.
Agent Cost Allocator MCP — multi-tenant LLM cost attribution for chargeback billing. Companion to
MCP gateway: one connection for your AI apps, with tool policies, PII redaction and audit logs.