MCP server for managing and checking ProxySEO proxy services. Provides tools to get proxy credentials, list assigned IPs, check proxy status, and verify proxy connectivity.
Four tools with clear verb-noun naming (get_proxy, list_proxies, proxy_status, check_proxy) and complete input schemas. All parameters have type and description. Tool descriptions are present (avg 95 chars) but lack actionable context for LLM selection, they state WHAT but not WHEN or WHY. No output schemas documented; responses are free-text JSON requiring LLM parsing. Error handling returns plain error strings without recovery guidance. Missing parameter constraints (e.g., IP format validation, port ranges). No mention of idempotency guarantees or retry safety despite idempotentHint annotations.
Verify that a proxy is working (connectivity + anonymity check through the proxy).
Get a proxy ready to use (IP:port + credentials). Returns connection URL or JSON.
List all proxy IPs assigned to your service.
Get the status of your proxy service (plan, auth type, IP count, ports).
No output schemas documented. Tools return free-text JSON responses (e.g., get_proxy returns map[string]interface{} serialized to JSON). LLMs cannot plan downstream calls or extract typed fields without knowing response structure.
Descriptions lack actionable context. 'Get a proxy ready to use' does not explain WHEN to call this vs list_proxies, or what prerequisites exist (e.g., must have assigned IPs). LLMs cannot disambiguate similar tools.
Error responses are plain strings ('Failed to get addresses: %v') without recovery guidance. LLMs cannot determine if the error is retryable, user-fixable, or fatal.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 53 | 2026-07-28+ | v2 |
check_proxy parameter 'ip' lacks format constraints. Description says 'Specific proxy IP to check' but does not specify IPv4 format, validation rules, or what happens if an invalid IP is passed.
Tool descriptions under 100 chars lack sufficient context. Baseline for A+ tools is 194 chars avg. Current descriptions (55 - 95 chars) are too terse to guide LLM selection in multi-tool scenarios.