Loading ToolBench

Reading the current corpus…

brainkb_mcp MCP server · F · ToolBench

brainkb_mcp

Exposes the BrainKB query_service API as MCP tools so an assistant can authenticate with credentials, ingest knowledge-graph data, read/search graphs, query W3C PROV-O provenance, and check ingest job status.

2026-07-28+STDIO15 tools

48/100

Definition Quality57

Protocol Readiness50

Supportability30

Spec posture

Inferred effective spec 2026-07-28+·current baseline 2026-07-28

Static source inference · medium confidence · evidence: stateless requests

Current-spec patterns detected

Summary

BrainKB MCP has 15 tools with consistent naming (verb_noun pattern: brainkb_*) and descriptions present for all tools. However, parameter descriptions are minimal or absent for most tools. Input schemas are visible but lack detail: most parameters have only type and description, missing constraints like enums, min/max bounds, or format specifications. Output schemas are not documented. Error handling guidance is absent. The server handles sensitive operations (login, token management, ingest) but lacks explicit permission gates and audit trail documentation.

Tools (15)

brainkb_create_tokenwriteauthsource verified67/100

Create a new Personal Access Token for the authenticated user.

brainkb_get_graphread onlyauth50/100

Get metadata and statistics for a specific knowledge graph.

brainkb_get_ingest_jobread onlyauth50/100

Get the status and details of a specific ingest job.

brainkb_globus_loginwritesource verified65/100

Authenticate with BrainKB using Globus OAuth.

brainkb_ingest_fileswriteauthsource verified73/100

Ingest RDF files from the server filesystem into a knowledge graph.

brainkb_ingest_rdfwriteauth50/100

Ingest RDF/Turtle data into a knowledge graph.

brainkb_list_graphsread onlyauth50/100

List all knowledge graphs accessible to the authenticated user.

brainkb_list_ingest_jobs

Issues (5)

HIGHtool ↗

Output schemas not documented. LLMs cannot infer what fields to expect from tool responses, forcing them to guess at downstream field names and risk broken tool chains.

brainkb_list_graphsbrainkb_get_graphbrainkb_search_graphsbrainkb_query_sparqlbrainkb_list_ingest_jobs

HIGHconstrained-input ↗

Parameter constraints missing. No enums for format (turtle|rdfxml|ntriples|jsonld), no min/max for limit/offset, no regex patterns for token/graph_id. LLMs will hallucinate invalid values.

brainkb_list_graphsbrainkb_search_graphsbrainkb_ingest_rdfbrainkb_ingest_files

HIGHrecovery-guide ↗

No error handling guidance. Tools lack recovery hints (e.g., 'If authentication fails, call brainkb_login first'). Agents cannot self-correct on failures.

brainkb_loginbrainkb_globus_loginbrainkb_query_sparql

Recommendations

Document output schemas for all tools. For brainkb_list_graphs, specify: {graphs: [{id, name, description, triple_count, created_at}], total_count, next_offset}. For brainkb_query_sparql, specify: {results: [{bindings: {var: value}}], query_time_ms}.
Add enum constraints to format parameter in brainkb_ingest_rdf: enum: ['turtle', 'rdfxml', 'ntriples', 'jsonld']. Add min/max to limit (1-1000) and offset (0+) in list/search tools.
Add error recovery hints to descriptions. E.g., brainkb_query_sparql: 'If query fails with 401, call brainkb_login or brainkb_use_token first. If syntax error, check SPARQL grammar.'
Document permission requirements for each tool. Add to descriptions: 'Requires: read:graphs' (brainkb_list_graphs) or 'Requires: write:ingest' (brainkb_ingest_rdf). Enable least-privilege agent configs.
Specify pagination behavior explicitly. For brainkb_list_graphs, add to description: 'Returns up to limit results. Use offset to paginate. Response includes total_count for planning.' Return total_count in response.

Score history

Overall score trend

First recorded score · v2 rubric

48/100

Scored	Grade	Overall	Spec posture	Rubric
2026-09-23	F	48	2026-07-28+	v2