MCP server for searching and surfacing Claude Code conversation history
Server has reasonable naming, complete parameter descriptions, and Zod-based schema definitions for all 9 tools. However, output schemas are entirely undocumented, tools return text content with no structured response format specified. Error handling is minimal; most tools return unstructured text. Cloud sync tools lack explicit idempotent/destructive annotations. Parameter naming is inconsistent (e.g., error_or_problem vs query). The server follows the basic pattern of verb+noun naming and has solid input validation coverage, but falls short of production-grade tool design due to missing output documentation and weak error guidance.
Pull knowledge entries and session summaries from the cloud server. Merges with local data without duplicates.
Push local knowledge entries and session summaries to the cloud server. Only pushes entries newer than the last sync.
Check cloud sync configuration and connection status. Shows whether sync is configured, last sync times, and connection health.
Discover recurring patterns in conversation history: common topics, frequent workflows, repeated issues.
Search past conversations for solutions to errors or problems. Prioritizes results containing fix/resolution language.
Get comprehensive context for a project: recent sessions, key topics, common tools, and patterns. Useful for session-start context injection.
Get a structured summary of a conversation session. Provide either a session ID or project name (returns most recent session).
Output schemas completely undocumented. All tools return {content: [{type: 'text', text: result}]} but the structure and fields within 'result' are never defined. LLMs cannot plan downstream operations or extract structured data.
Error handling is minimal. Most tools return unstructured text responses with no recovery guidance. When search fails, invalid query is provided, or cloud sync fails, LLMs get generic text with no indication of what to do next.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-22 | F | 0 | 2026-07-28+ | v2 |
| 2026-03-09 | C | 61 | - | v1 |
List all projects that have Claude Code conversation history, with session counts and activity dates.
Search past Claude Code conversations. Supports filter syntax: project:name, before:date, after:date, tool:name. Dates can be relative (7d, 30d, 1w) or ISO (2024-01-15).
Destructive operations (cloud_sync_push, cloud_sync_pull with force=true) lack explicit safety annotations or confirmation mechanisms. No dry-run option. LLMs could trigger unwanted data synchronization without warning.
Parameter naming inconsistency. 'search_history' uses 'query' but 'find_solutions' uses 'error_or_problem'. 'find_solutions' uses 'technology' (string) while 'search_history' has 'tool' in description. Inconsistent naming across similar tools confuses LLMs.
No pagination or result limiting guidance in tool descriptions. Tools like 'get_project_context' and 'find_patterns' could return large result sets, but descriptions don't mention limits or how to constrain results.