Loading ToolBench

Reading the current corpus…

mcp-upgrade MCP server · F · ToolBench

← Browse

mcp-upgrade

github.com/JacobRyan258/mcp-upgrade

Find what breaks before you upgrade your MCP server. Static scanner for MCP 2025-11-25 -> 2026-07-28 RC migration.

<=2025-11-25STDIO4 tools

31/100

Definition Quality29

Protocol Readiness16

Supportability45

Spec posture

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

Static source inference · medium confidence · detected: Roots, Logging

Deprecated protocol patterns detected

Summary

This is a test fixture repository, not a production MCP server. The 4 tools have minimal definitions: 'echo' and 'readme' lack input schemas entirely; 'lookup' and 'render_report' have only single string parameters with no descriptions. All tool descriptions are trivial (10-20 chars), well below the 50-200 char baseline for LLM-optimized tools. No output schemas documented. No error handling guidance. The codebase is a scanner/validator for MCP servers, not a functional MCP server itself. Scoring reflects the fixture-quality definitions visible in test files.

Tools (4)

echoread onlysource verified38/100

Returns its input unchanged.

lookupread onlysource verified27/100

Lookup

readmeread onlysource verified17/100

Readme resource

render_reportread onlysource verified32/100

Renders a large report.

Issues (5)

CRITICALtool ↗

Tool 'readme' has no input schema visible in source.

readme

CRITICALtool-description ↗

All tool descriptions are under 20 characters (e.g. 'Readme resource', 'Lookup'). Descriptions must be 50-200 chars and explain WHAT, WHEN, and WHY.

echoreadmelookuprender_report

HIGHtool-description ↗

No parameter descriptions provided for any tool. 'id' in lookup and 'reportId' in render_report lack context. LLMs cannot infer whether these are UUIDs, integers, or slugs.

lookuprender_report

HIGHtool ↗

No output schemas documented for any tool. LLMs cannot plan downstream calls or extract required fields.

echoreadmelookuprender_report

HIGH

Recommendations

Expand all tool descriptions to 50-200 characters. Example: 'lookup' → 'Look up a resource by its unique ID. Returns the resource metadata and content. If not found, returns available IDs.'
Add descriptions to every parameter. Example: 'id' → 'The unique identifier of the resource (UUID format, e.g. 550e8400-e29b-41d4-a716-446655440000).'
Document output schemas for each tool. Specify field names, types, and what each field represents.
Add error handling guidance: 'If lookup fails with 404, call list_resources() to discover valid IDs.'
For 'render_report', clarify the expected format of reportId and what the output contains (HTML, PDF, JSON, etc.).

Spec posture evidence

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

Relies on Roots (deprecated) - pass files/dirs via tool parameters, resource URIs, or configuration

Score history

Overall score trend

First recorded score · v2 rubric

31/100

Scored	Grade	Overall	Spec posture	Rubric
2026-09-23	F	31	<=2025-11-25	v2