Loading ToolBench
A complete, production-grade Model Context Protocol server for the Render Public API — all 208 official endpoints plus higher-level workflow tools.
Strong foundation with 6 tools covering Render API operations and self-discovery. All tools have clear descriptions (50-200 chars range) and well-structured input schemas with type definitions and constraints. Naming follows verb_noun pattern consistently. Composite tools (render_find_service, render_wait_for_deploy, render_service_status) demonstrate thoughtful composition for common workflows. However, output schemas are not explicitly documented in the visible code, and error handling guidance is minimal. Tool annotations present but limited to read-only hints.
Returns the command or configuration that installs the render-useful-mcp MCP server for a given client (Claude Code, Claude Desktop, VS Code, Cursor, or any client using an mcpServers object). Claude Code commands install globally, into user scope.
Returns a complete, copyable example of a common Render task performed through render-useful-mcp tools — creating a Docker cron job, deploying and waiting, diagnosing a broken service, or reading logs — including the exact tool name and argument shape.
Resolves a service name — including a partial or approximate one — to a single Render service, returning its id plus close alternatives. Use this first whenever the user names a service instead of giving an id; every other service tool needs the id.
One-call triage for a service: its configuration, latest deploys, running instances and most recent error-level logs. Use this to answer "is X healthy?" or "why is X broken?" instead of chaining retrieve/list/logs calls.
Polls a deploy until it reaches a terminal state (live, build_failed, update_failed, canceled, deactivated) or the timeout expires. Defaults to the service's most recent deploy. Use after render_create_deploy instead of polling render_retrieve_deploy in a loop.
Output schemas not documented in visible code. LLMs cannot plan downstream tool calls or extract fields without knowing response structure.
Error handling lacks recovery guidance. No indication of which errors are retryable, user-fixable, or fatal. Responses for 204 status return generic success message instead of context-specific feedback.
Tool annotations minimal: only readOnlyHint present. Missing destructiveHint and idempotentHint for tools that modify state (render_create_deploy implied but not visible in sample).
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | C | 63 | <=2025-11-25 | v2 |
Searches the tools that render-useful-mcp exposes for the Render Public API, by name, toolset or description. Use it to answer which tool performs a given Render operation — creating a service or cron job, reading logs, scaling, managing Postgres. Returns tool names with their descriptions and whether they are read-only or destructive.
Composite tools (render_find_service, render_wait_for_deploy) accept both service name and ID but lack explicit documentation of resolution order or fallback behavior when both are provided.
search_render_mcp_tools and get_render_mcp_recipe use enum constraints for parameters but lack examples in descriptions of what results look like or when to use each task variant.