Loading ToolBench
MCP server for F5 Distributed Cloud API that exposes F5XC APIs to AI assistants via Model Context Protocol. Supports dual-mode operation: documentation mode (unauthenticated) and execution mode (authenticated with F5XC credentials).
The F5XC API MCP server provides 13 tools with generally well-structured schemas and descriptions. Most tools have clear naming conventions (search_, execute_, describe_) and detailed parameter documentation. However, there are several quality gaps that prevent a higher score: (1) Missing output schema documentation for all tools, while input schemas are present and well-described, return types are not formally specified; (2) All tools are READ_ONLY or WRITE risk classification, but descriptions lack explicit state-change warnings for destructive operations; (3) Some parameter descriptions contain examples inline (e.g., 'f5xc-api-waap-http-loadbalancer-create') which LLMs may reuse literally; (4) No tool explicitly documents error recovery paths or what to do if a call fails; (5) Tools 11-13 (quota-limits, quota-status, quota-reset) have identical empty input schemas with no documented return structure, making them appear incomplete. The discovery and execution tools are well-designed and follow composition patterns (search → describe → execute workflow), but the suite lacks the output documentation and error-handling guidance expected of A-grade tools.
Get domain-specific best practices, common errors, and recommended workflows for F5XC resource management.
Analyze resource dependencies and creation ordering. Returns prerequisite resources, dependent resources, and recommended creation sequence.
Get detailed information about a specific F5XC API tool including parameters, request body schema, and usage examples. Use search_tools first to find the tool name.
Estimate token and latency costs for tool execution. Helps plan workflows and understand performance implications.
Execute a CRUD operation on a consolidated F5XC resource. Specify the resource name and operation. Routes to the correct underlying API tool automatically.
Execute an F5XC API tool. In authenticated mode, makes the actual API call. In documentation mode, returns CLI equivalents and curl examples.
Missing output schema documentation for all tools. While input parameters are well-documented, there is no formal specification of what each tool returns (response fields, types, structure). LLMs cannot plan downstream tool chains without knowing available fields.
Tools 11-13 (f5xc-api-quota-limits, f5xc-api-quota-status, f5xc-api-quota-reset) have empty input schemas (no properties defined) with minimal descriptions (40 chars). Descriptions lack detail on what quota information is returned, formats, or refresh intervals.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-22 | D | 54 | <=2025-11-25 | v2 |
| 2026-03-09 | F | 19 | 1.0.0+ | v1 |
Check API rate limits and quota information for authenticated user.
Get quota reset schedule and renewal information.
Get current quota usage and remaining capacity.
Search for F5XC resources (consolidated view). Returns resources with their available CRUD operations. More efficient than searching individual tools - one result per resource instead of 5 CRUD tools.
Search for F5XC API tools using natural language. Returns matching tools with relevance scores. Use this to find tools for specific operations like 'create load balancer' or 'list DNS zones'.
Get F5XC API MCP server information including authentication status, available domains, and total tool count. Use this to understand server capabilities.
Validate tool parameters before execution. Checks for required fields, type mismatches, and constraint violations based on OpenAPI schemas.
Descriptions for execute tools (f5xc-api-execute-tool, f5xc-api-execute-resource) lack explicit warning that these are state-changing operations. Agents should understand destructive vs idempotent calls. Descriptions do not state 'this modifies resources' or 'use with caution'.
Parameter descriptions contain example tool names inline (e.g., 'f5xc-api-waap-http-loadbalancer-create', 'f5xc-api-dns-zone-update'). LLMs frequently reuse example values literally in real API calls. Examples should be in enums or pattern constraints, not free-text descriptions.
No error handling guidance in any tool descriptions. Tools lack explicit recovery paths (e.g., 'If tool not found, try f5xc-api-search-tools() first' or 'If validation fails, check parameter types'). Agents cannot self-correct on failures without explicit guidance.
Tool f5xc-api-estimate-cost has ambiguous parameter structure: 'tools' array is documented alongside 'toolName' string. Descriptions do not clarify if both are optional, mutually exclusive, or both required. Parameter relationship is undocumented.
Tools f5xc-api-search-tools and f5xc-api-search-resources accept 'limit' with default 10 but no documented maximum. Baseline rubric requires min/max bounds for numeric parameters to prevent LLMs from requesting thousands of results.