Skip to main content

Security Limits

Use this page for safety guardrails on Gateway MCP usage.

Guardrail categories​

1) Scope limits​

  • Least-privilege profile for each agent
  • Restrict allowed methods/routes via profile policy
  • Use separate identities per environment

2) Throughput limits​

  • Profile rate limits
  • Quota ceilings for sustained traffic
  • Ingress concurrency controls for burst protection

3) Blast-radius limits​

  • Block destructive tools/routes unless explicitly required
  • Keep production write scope isolated from read-only automation

Validation sequence​

  1. Allowed call with expected profile -> success.
  2. Out-of-scope call -> 403.
  3. Missing auth -> 401.
  4. Burst run above limits -> 429.
  5. Confirm request logs show deny reason and client/profile identity.

Example limit test​

curl -i https://gateway.example.com/mcp \
  -H "Authorization: Bearer $MCP_TOKEN" \
  -H "X-Client-ID: ai-agent-prod" \
  -H "X-Profile-ID: mcp-prod" \
  -H "Content-Type: application/json" \
  -d '{"jsonrpc":"2.0","id":"tools-1","method":"tools/list","params":{}}'