MiniMax-AI / MiniMax-AI/MiniMax-Coding-Plan-MCP

[BadCase about the model]: MiniMax M2 sometime doesn't pick the result of the MCP (in Cursor)

Open
#10 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Python
Stars
102
Forks
30
PR merge metrics
No merged PRs in 30d

Description

Basic Information - Models Used

MiniMax-M2

Basic Information - Scenario Description

MiniMax M2 doesn't pick the result of MCP after it's being called:

Is this badcase known and solvable?
Information about environment

OS: MAC Tahoe 26.1
Python: 3.12
IDE: Cursor Version: 2.2.0-pre.31.patch.0

Call & Execution Information

Call type: MCP call via Cursor

What it send:
{
"prompt": "Please analyze this image and describe what you see. What is the main content, visual elements, text, or any important details?",
"image_source": "https://pbs.twimg.com/media/G7lGwQCa8AAk2hl?format=jpg&name=large"
}

Description
Steps to reproduce

The bug can be reproduced with the following steps:

  1. use Cursor the nightly version
  2. use this prompt
    help me to understand about this image https://pbs.twimg.com/media/G7lGwQCa8AAk2hl?format=jpg&name=large
Expected behavior

The results are expected to be: after receiving the result tools, it should follow the result.

Actual behavior

The actual outputs are as follows: after the tools complete, the model might didn't pick it up

Image

and event it trying to download it, which is wrong.

I'm testing with other model like Opus 4.5 or sonnet 4.5 , it can utilize the MCP of minimax coding correctly.

Error logs

The error logs are as follows:

No Error log

Contributor guide

No contributing guide indexed for this repository

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start by reproducing the MCP call in Cursor using the listed image prompt and environment, then compare the tool result with the model's subsequent response. The issue names no repository files or tests, so identify the MCP entry point handling the call before determining what a correct result handoff should be. Done means the model consistently uses the completed tool result rather than attempting to download the image.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
api
Issue type
Bug
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.