MCP Client Compatibility Probe
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| MCP_PROBE_TRACE | No | Path to the trace log file (NDJSON). Defaults to /tmp/mcp-probe.ndjson. | /tmp/mcp-probe.ndjson |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": true
} |
| prompts | {
"listChanged": true
} |
| resources | {
"subscribe": false,
"listChanged": true
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| echo_metaA | Returns the received arguments and MCP request metadata. Use this first to inspect protocolVersion, clientInfo, and clientCapabilities. |
| structured_resultB | Returns both text content and structuredContent conforming to outputSchema. |
| create_handleA | Creates an explicit short-lived probe handle to test stateless multi-call tool design. |
| use_handleC | Uses a handle returned by create_handle. Unknown handles return a tool execution error. |
| needs_form_inputA | Returns resultType input_required until the client retries with inputResponses. This tests MRTR and elicitation form mode. |
| tool_errorA | Always returns a tool execution error via isError true, not a JSON-RPC protocol error. |
| resource_link_resultC | Returns a resource_link content item pointing at a probe resource. |
| searchA | ChatGPT-compatible read-only search stub. Returns result IDs, titles, and URLs. Use fetch to retrieve full text. |
| fetchA | ChatGPT-compatible read-only fetch stub. Retrieves full text for an ID returned by search. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
| probe_summary | Prompt template that asks the model to summarize observed MCP client behavior. |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
| overview | Short overview resource exposed by the probe server. |
| client-observation | Checklist for interpreting the probe trace log. |
TDQS
Scored across 9 tools
Each tool targets a distinct MCP feature: metadata inspection, structured output, handle-based state, MRTR input, error simulation, resource links, and search/fetch stubs. There is no overlap between their purposes.
Names are a mix of verb_noun (echo_meta, create_handle), standalone verbs (search, fetch), and nouns (structured_result, tool_error). While all are snake_case, the varying forms make the naming pattern less predictable than a uniform verb_noun convention.
9 tools is a well-scoped set for a compatibility probe, covering the key MCP client interaction patterns without redundancy or bloat.
The tool surface covers essential probe scenarios: metadata, structured content, handles, MRTR, errors, resource links, and search/fetch. Minor gaps exist (e.g., no explicit tool for protocol-level logging or sampling), but the core compatibility checks are well represented.