Loading ToolBench

Reading the current corpus…

sluice MCP server · F · ToolBench

← Browse

sluice

github.com/nnemirovsky/sluice

A SOCKS5 proxy server with policy enforcement, credential injection, approval workflows, and MCP gateway capabilities

<=2025-11-25STREAMABLE_HTTP4 tools

32/100

Definition Quality25

Protocol Readiness42

Supportability36

Spec posture

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

Static source inference · medium confidence · detected: stateful session

Deprecated protocol patterns detected

Summary

Sluice is an HTTP gateway/proxy for upstream MCP servers. Tool definitions are minimal and inferred from test code rather than explicit registration. The 4 'tools' listed (tools/list, tools/call, initialize, notifications/initialized) are not standard MCP tools but rather HTTP endpoints or protocol messages. No input schemas are visible in the source. Descriptions are generic and under 50 characters. The server acts as a multiplexer, not a tool provider, it proxies calls to upstream servers. This architectural mismatch means the tool definitions do not follow MCP patterns for agent-facing tools.

Tools (4)

initializeread onlysource verified23/100

Initialize an MCP session with protocol version and client capabilities

notifications/initializedread only20/100

Notification sent by client after initialization is complete

tools/callwriteauthsource verified30/100

Call a tool from an upstream MCP server with provided arguments

tools/listread onlyauth27/100

List all available tools from upstream MCP servers

Issues (5)

CRITICALtool ↗

No input schemas visible for any tool. All schema scores are 0. Cannot verify parameter types, constraints, or descriptions.

tools/listtools/callinitializenotifications/initialized

CRITICALtool ↗

Tool definitions are inferred from test code (server_http_test.go) rather than explicit registration.

tools/listtools/callinitializenotifications/initialized

HIGHtool-description ↗

Descriptions are generic and under 50 characters. 'List all available tools from upstream MCP servers' (47 chars) lacks context on when/why to call or what structure is returned.

tools/listtools/call

HIGHconstrained-input ↗

tools/call accepts 'arguments' as a free-form object with no schema constraints. LLMs cannot infer valid argument structure for upstream tools.

tools/call

Recommendations

Add explicit JSON Schema definitions for all tool inputs. For tools/call, define a schema that validates the 'arguments' object structure based on the upstream tool's schema.
Expand tool descriptions to 100-200 characters. Include WHAT the tool does, WHEN to use it, and WHAT it returns. E.g., 'List all tools available from connected upstream MCP servers. Returns tool name, description, and input schema for each. Call this first to discover available operations.'
Document the output schema for each tool. Specify field names, types, and whether results are paginated. E.g., tools/list should return {tools: [{name, description, inputSchema}], total: number}.
Add error handling that categorizes failures as retryable (network timeout), user-fixable (invalid tool name), or fatal (upstream server crashed). Include recovery guidance in error messages.
Implement per-request _meta with logLevel support (current spec 2026-07-28) to enable stateless request handling and per-request logging configuration.

Spec posture evidence

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

Relies on Stateful initialize / Mcp-Session-Id (removed; protocol is stateless) - make each request self-contained

Score history

Overall score trend

First recorded score · v2 rubric

32/100

Scored	Grade	Overall	Spec posture	Rubric
2026-09-23	F	32	<=2025-11-25	v2