Loading ToolBench

Reading the current corpus…

notion-mcp MCP server · D · ToolBench

← Browse

notion-mcp

github.com/gwicho38/notion-mcp

MCP server that caches Notion blog posts locally with FTS5 search

<=2025-11-25STDIO6 tools

55/100

Definition Quality79

Protocol Readiness47

Supportability21

Spec posture

Inferred effective spec <=2025-11-25·current baseline 2026-07-28

Static source inference · medium confidence · detected: Logging

Deprecated protocol patterns detected

Summary

Notion MCP demonstrates solid tool design with complete schemas, clear descriptions, and proper tool annotations. All 6 tools have input schemas with typed parameters and descriptions. Tool names follow verb_noun convention (blog_ingest, blog_search, blog_delete). Descriptions are substantive (100-200 chars typical). However, output schemas are not formally documented, responses are JSON strings without declared structure. Error handling is minimal (only 'Cache not initialized' check visible). No parameter validation beyond Pydantic constraints. Missing guidance on recovery paths for common failures.

Tools (6)

blog_deletedestructivesource verified78/100

Remove a blog post from the local cache.

blog_getread onlysource verified78/100

Retrieve the full markdown content of a cached blog post.

blog_ingestwritesource verified80/100

Ingest a Notion blog post into the local cache. Call this after reading a blog post from the Notion MCP. The post content is cached locally for fast search and retrieval.

blog_ingest_batchwritesource verified78/100

Ingest multiple blog posts at once for bulk sync.

blog_listread onlysource verified78/100

List all cached blog posts, optionally filtered by category or status.

blog_searchread onlysource verified83/100

BM25 full-text search across cached blog posts. Searches title, content, and categories. Returns ranked results with relevance scores and text snippets.

Issues (5)

HIGHresponse-shaper ↗

Output schemas not formally documented. Tools return JSON strings without declared field structure. LLMs cannot plan downstream calls or extract fields reliably.

blog_ingestblog_ingest_batchblog_deleteblog_searchblog_listblog_get

HIGHrecovery-guide ↗

Error handling lacks recovery guidance. 'Cache not initialized' is the only visible error path. No guidance on retryability, user-fixable errors, or next steps.

blog_ingestblog_ingest_batchblog_deleteblog_searchblog_listblog_get

MEDIUMpaginated-result ↗

blog_search and blog_list lack pagination parameters (limit, offset, cursor). Large result sets could exhaust context window.

blog_searchblog_list

MEDIUMconfirmation-request ↗

Recommendations

Document output schemas for all tools. Define field names, types, and meanings. E.g., blog_search should declare: {results: [{page_id, title, relevance_score, snippet}], total_count}
Add error classification and recovery guidance. E.g., 'Cache not initialized (fatal): restart the server' or 'Post not found (user-fixable): try blog_search() to find the correct page_id'
Add pagination to blog_search and blog_list. Include limit (default 10, max 50), offset/cursor, and total_count in responses.
Implement dry-run for blog_delete. Add optional 'dry_run: bool' parameter; return what would be deleted without actually deleting.
Enhance parameter descriptions with format constraints. E.g., 'Publication date in ISO 8601 format (YYYY-MM-DD)' instead of just '(ISO format)'

Spec posture evidence

Inferred effective spec: <=2025-11-25.

Relies on Logging (deprecated) - log to stderr or use OpenTelemetry

Score history

Overall score trend

First recorded score · v2 rubric

55/100

Scored	Grade	Overall	Spec posture	Rubric
2026-09-23	D	55	<=2025-11-25	v2