Click on "Install Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@MCP-Grounded classify this skin lesion image, abstain if uncertain"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
A multi-agent pipeline for medical image classification with verification-aware abstention, coordinated via the Model Context Protocol (MCP).
"Instead of always guessing, the AI says β I'm not confident enough, I'll skip this one."
MCP-Grounded is a 4-agent AI pipeline that classifies skin lesion images from the HAM10000 dataset. What makes it novel: the final agent can abstain from answering when it isn't confident β making it safer for medical use.
All four agents are real MCP tools, not just described as such.
Related MCP server: Prior Authorization System MCP Server
Skin lesion image
β
βΌ
βββββββββββββββββββ
β BiomedCLIP β Agent 1: Extract 512-dim embedding
β (Extract) β
ββββββββββ¬βββββββββ
β
βΌ βββββββββββββββββββββββββββββββββββββββββββ
β MCP Server β
β β
β ββββββββββββ ββββββββββββ β
β β Retrieve ββββββββΆβ Rerank β β
β β Agent 2 β β Agent 3 β β
β ββββββββββββ ββββββ¬ββββββ β
β β β
β ββββββββββΌβββββββββββ β
β β Verify / Abstain β β
β β Agent 4 β β
β ββββββββββ¬ββββββββββββ β
βββββββββββββββββββββββββββ β ββββββββββββββ
β
βββββββββββββββ΄ββββββββββββββ
β β
conf β₯ Ο conf < Ο
β β
PREDICT ABSTAINMetric | Value |
Recall@1 | 77.9% |
Recall@5 | 93.5% |
Recall@10 | 96.3% |
Recall@50 | 99.2% |
Threshold Ο | Coverage | Selective Accuracy |
0.0 (answer all) | 100.0% | 67.0% |
0.5 | 96.9% | 69.0% |
0.6 | 83.4% | 77.0% |
0.7 | 52.0% | 91.3% |
0.8 | 4.9% | 98.6% |
At Ο = 0.7, selective accuracy improves +24 percentage points over the no-abstention baseline.
As the confidence threshold rises, coverage drops but selective accuracy climbs sharply β proving abstention makes the system safer.
Metric | Value |
ECE before temperature scaling | 0.191 |
ECE after temperature scaling | 0.185 |
Learned temperature T | 0.944 |
HAM10000 β 10,015 dermoscopic images across 7 skin lesion categories:
akiec Β· bcc Β· bkl Β· df Β· mel Β· nv Β· vasc
Split: 70% train / 15% validation / 15% test (stratified).
Open notebook1_embeddings.py in Google Colab with a T4 GPU runtime. Run all cells top to bottom. Downloads HAM10000 and produces embeddings.npz.
Open notebook2_experiments.py in a new Colab notebook. Upload embeddings.npz. Run all cells. Produces:
All result tables (Recall@K, accuracy, calibration, abstention)
risk_coverage.png
clf_weights.npz
pip install "mcp[cli]" numpy torch
python mcp_grounded_server.pyStarts a live MCP server with three callable tools: retrieve, rerank, classify_and_verify.
mcp[cli]
numpy
torch
open_clip_torch
scikit-learn
pandas
pillow
tqdm
matplotlibSee requirements.txt.
mcp_grounded/
βββ notebook1_embeddings.py # Colab: download HAM10000, extract BiomedCLIP embeddings
βββ notebook2_experiments.py # Colab: retrieval, calibration, abstention experiments
βββ mcp_grounded_server.py # Local: FastMCP server exposing 4 agents as tools
βββ risk_coverage.png # Figure 2: risk-coverage curve
βββ requirements.txt
βββ README.mdIf you use this work, please cite:
@inproceedings{mcpgrounded2025,
title = {MCP-Grounded: A Multi-Agent Pipeline with Verification-Aware Abstention for Medical Image Classification},
author = {[Your Name]},
booktitle = {[Conference Name]},
year = {2025}
}MIT License. Dataset (HAM10000) is CC-BY-NC-SA-4.0 β see Kaggle for terms.
Built with BiomedCLIP Β· FastMCP Β· HAM10000
This server cannot be installed
Unclaimed servers have limited discoverability.
If you are the server author, to access and configure the admin panel.
MERCATOR Verify: evidence-backed verification and decision support for autonomous agents.
Trust gate for AI agents: multi-model adversarial consensus, signed and verifiable verdicts.
AI visual generation agent: multi-pipeline rendering, prompt crafting, and image composition.
Multi-agent governance: task orchestration, compliance, decision validation, and ML predictions.
