A Model Context Protocol server for remote Windows crash dump analysis using WinDbg/CDB
Server has 7 tools with complete input schemas and descriptions. Naming follows verb_noun convention (prepare_, start_, execute_, get_, close_). Descriptions are present but brief (avg 60 chars, below 194-char baseline). Parameters are typed and described, but lack constraints (enums, ranges, patterns). Output schemas are not documented in the visible code. Error handling exists but recovery guidance is minimal. No tool annotations (readOnlyHint/destructiveHint) despite clear risk classifications in metadata.
Close an analysis session and clean up resources
Execute a CDB command synchronously in an analysis session
Get the result of an asynchronously executing CDB command
Get the status of an asynchronously executing CDB command
Prepare a dump file upload session
Start a crash dump analysis session
Start a CDB command asynchronously in an analysis session
Output schemas not documented. Tools return results but LLMs cannot infer response structure, forcing guesswork on field names and types.
Tool descriptions are too brief (avg 60 chars vs 194-char baseline). Missing WHEN to use each tool and dependencies between them (e.g., prepare_dump_upload must precede start_analysis_session).
No tool annotations despite clear risk classifications. execute_windbg_command and async variants are READ_ONLY but lack readOnlyHint. prepare_dump_upload, start_analysis_session, close_analysis_session are WRITE but lack destructiveHint.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | C | 60 | <=2025-11-25 | v2 |
Parameter constraints missing. 'timeout' in execute_windbg_command has no min/max bounds (default 1800s). 'command' parameters accept free-form strings with no validation hints for CDB syntax.
Error handling lacks recovery guidance. No indication of which errors are retryable, which require user intervention, or what the LLM should do next.