5ire is a desktop MCP client, not an MCP server providing tools to agents. The 8 'tools' listed are actually internal service functions for managing MCP server configurations and tool invocation, not tools that agents would call. Tool definitions lack proper descriptions, parameter validation, error guidance, and output schemas. Most tools have minimal documentation (under 50 chars) and incomplete schemas. The 'call' tool is particularly problematic: it exposes raw MCP tool invocation without safety rails, error recovery guidance, or output structure documentation. Schema quality is inconsistent, some tools have object-typed 'options' params with nested properties, but these lack type declarations and descriptions for nested fields. No tool provides enums, ranges, or format constraints. Error handling is absent from descriptions. The liveServers and liveActiveServers tools document behavior ('continuously pushes updates') but lack schema definitions for return types.
Activates an existing MCP server in the database
Calls a tool from an MCP server using the provided URI and input parameters. The URI format must follow: tool:{connection-id}/{tool-name}
Creates a new MCP server in the database
Deactivates an existing MCP server in the database
Deletes an existing MCP server from the database
Listen to active server changes in real-time. Returns an object with subscribe method, refresh function and initial results that continuously pushes updates of active servers
Listen to server changes in real-time. Returns an object with subscribe method, refresh function and initial results that continuously pushes updates of servers
No output schemas documented for any tool. LLMs cannot predict return types, fields, or structure. liveServers and liveActiveServers claim to return objects with 'subscribe method, refresh function, and initial results' but no schema validates this.
Parameter descriptions are minimal or absent. 'options' parameter in createServer/updateServer/deleteServer/activateServer/deactivateServer is documented only as 'The options for [action]'. Nested properties like 'transport', 'endpoint', 'projectId' lack individual descriptions explaining their purpose, constraints, and valid values.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-22 | F | 42 | 2025-06-18+ | v2 |
| 2026-03-09 | F | 0 | 1.17.1+ | v1 |
Updates an existing MCP server in the database
Nested schema properties lack type declarations. In createServer.options, properties 'transport', 'endpoint', 'projectId' are listed but have no 'type' field. updateServer.options.config is typed 'object' with no property definitions. Schemas violate JSON Schema conventions.
No enum constraints on 'transport' parameter. createServer accepts 'transport' as free-form string when it should be enum ['stdio', 'http']. Allows hallucinated invalid values like 'websocket' or 'grpc'.
No error handling guidance in descriptions. createServer, updateServer, deleteServer, activateServer, deactivateServer lack any indication of what errors are possible, which are retryable, or what the LLM should do next if the call fails.
'call' tool exposes raw MCP tool invocation without safety guardrails. Description states it 'Calls a tool from an MCP server' but provides no guidance on error recovery, output handling, or limits. No documentation of what happens when the remote tool fails, times out, or returns invalid data.
Destructive operations (deleteServer) lack confirmation mechanism. Description does not indicate this is irreversible or suggest a dry-run option. An agent could delete a server without explicit user approval.
liveServers and liveActiveServers have no documented return schema. Description claims they return 'an object with subscribe method, refresh function and initial results' but provides no type definitions, field names, or data structure. Impossible for an LLM to extract data from these responses.
Tool descriptions do not clarify WHEN to use each tool or what the LLM should expect. activateServer and deactivateServer have nearly identical short descriptions with no indication of their effects or prerequisites.