This is a test/evaluation fixture server, not a production MCP server. Tools are defined across multiple Python test fixtures (echo_server.py, probe_broken_server.py, probe_clean_server.py, probe_paged_server.py, transport_fault_server.py). Most tools have minimal or single-character descriptions ('x', 'a', 'b', 's', 'g', 'd', 'retry', 'break', 'recover', 'validate', 'parallel', 'lookup'). Input schemas are present but sparse, most have empty properties. Three tools (ok, slow, crash) lack descriptions entirely. Parameter descriptions are absent across all tools. No output schemas documented. Error handling is not visible in tool definitions. This server appears designed to test MCP client robustness against malformed/broken servers, not to serve as a production tool reference.
a
b
break
d
status
retry
g
14 of 17 tools have descriptions under 20 characters or missing entirely (e.g., 'read', 'status', 'retry', 'break', 'recover', 'validate', 'parallel', 'a', 'b', 's', 'g', 'd', and three with no description).
No parameter descriptions across any tool. LLMs cannot infer what 'waitUntil', 'port', or 'key' parameters control without explicit descriptions.
No output schemas documented for any tool. LLMs cannot plan downstream calls or extract required fields without knowing what each tool returns.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 40 | 2025-06-18+ | v2 |
lookup
Navigate with optional wait condition
read
recover
validate
parallel
s
Tool names lack action verbs. 'navigate', 'lookup', 'shared_read' are acceptable, but 'read_counter', 'describe_status', 'alpha_read', 'beta_status', 'slow_read', 'gamma_reset', 'delta_slow' are vague or abbreviated. Single-letter names ('a', 'b', 's', 'g', 'd') are not self-documenting.
describe_status has a required parameter 'missing' that does not exist in the inputSchema properties, creating a schema validation error. This will cause tool invocation failures.