Loading ToolBench
MCP Flux Pro demonstrates good definition quality with clear, descriptive tool names and generally well-documented parameters. All 6 tools follow verb_noun naming conventions (flux_generate_image, flux_edit_image, flux_list_models, etc.). Tool descriptions are comprehensive and explain both WHAT the tools do and WHEN to use them, with explicit guidance on tool selection (e.g., 'For editing existing images, use flux_edit_image instead'). However, there are notable gaps: (1) output schemas are not explicitly documented in the source, descriptions mention 'Returns: Task ID and generated image information' but actual response structure is not visible, forcing LLMs to guess at output fields; (2) error handling guidance is absent, tools provide no recovery instructions for common failures; (3) tool annotations (readOnlyHint, destructiveHint) are missing despite some tools being write operations. The two generation/edit tools correctly identify themselves as WRITE risk, but the server does not use MCP's tool annotation patterns. Parameter descriptions are strong (60-100+ chars each, well-formatted examples), but lack explicit validation rules (e.g., size formats should state 'pixel dimensions must be multiples of 32'). The flux_list_* tools are well-designed discovery tools that guide users before calling generative tools. Overall, this is solid foundation-level work with room for output documentation and error patterns.
Edit an existing image using Flux with a text prompt. This allows you to modify an existing image based on a text description. The kontext models (flux-kontext-pro, flux-kontext-max) are specifically designed for high-quality image editing and style transfer. Use this when: - You want to modify or transform an existing image - You want to change specific elements in an image - You want to apply style changes or artistic effects - You want to add, remove, or replace objects in an image For generating new images from scratch, use flux_generate_image instead.
Generate AI images from a text prompt using Flux. Flux is a family of fast, high-quality image generation models by Black Forest Labs. Different models offer different tradeoffs between speed, quality, and capabilities. Use this when: - You want to create new images from a text description - You need high-quality AI-generated artwork or photos - You want fast image generation with good prompt following For editing existing images, use flux_edit_image instead.
Query the status and result of a Flux image generation task. Use this to check if a generation is complete and retrieve the resulting image URLs and metadata. Use this when: - You want to check if an image generation has completed - You need to retrieve image URLs from a previous generation - You used async callback and want to check results later - The initial generation returned a task_id without immediate results
Query multiple Flux image generation tasks at once. Efficiently check the status of multiple tasks in a single request. More efficient than calling flux_get_task multiple times. Use this when: - You have multiple pending generations to check - You want to get status of several images at once - You're tracking a batch of generations
Output schemas not documented. Tools return strings/objects but actual structure is not visible in source. LLMs cannot plan downstream calls or extract specific fields (e.g., task_id, image_urls). This violates the 'Document the output schema' pattern.
Error handling and recovery guidance missing. Tools provide no actionable error messages or next steps. E.g., if image generation fails, tool should say 'Generation failed due to invalid prompt, try removing unusual characters' or 'Task not found, check task_id or wait for async callback.' Current implementation likely returns opaque API errors.
Tool annotations missing. flux_generate_image and flux_edit_image are destructive WRITE operations that should declare destructiveHint=true and idempotentHint=false in MCP tool registration. Read-only tools (flux_list_*) should declare readOnlyHint=true. This metadata helps agents reason about side effects.
Inferred effective spec: 2026-07-28+.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-22 | B | 77 | 2026-07-28+ | v2 |
| 2026-03-09 | F | 49 | - | v1 |
List all available Flux tools and their use cases. Reference guide for what each tool does and when to use it.
List all available Flux models and their capabilities. Reference guide for choosing the right Flux model for your use case.
Parameter validation rules not explicit. Size parameter accepts 'pixel dimensions' or 'ratios' depending on model, but constraint text is vague. Should state: 'For flux-dev: exact pixels (multiples of 32, 256-1440px). For kontext: strict ratios only (1:1, 16:9, etc.).' Currently LLMs may pass invalid formats without clear rejection guidance.
count parameter in flux_generate_image states 'Only supported for generate action' but unclear if other tools ignore it silently or return errors. Should document: 'count is ignored for edit operations; use flux_generate_image for batch generation.'