supply-ops-agent
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., "@supply-ops-agentCheck the health of Partner Bravo"
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.
supply-ops-agent
A supply-side operations diagnostic agent built on AAMP standards, extending the IAB Tech Lab seller agent framework with diagnostic tooling for bid rate analysis, IVT detection, and ads.txt compliance. Built with MCP for integration with any AAMP-compatible client.
What it does
This MCP server exposes five diagnostic tools that help sell-side ad ops teams investigate supply health issues:
Tool | Description |
| Daily bid rates, anomaly detection vs baseline, win rate and volume |
| Declared vs observed sellers, mismatch flagging, detection timestamps |
| IVT rates broken down by GIVT/SIVT, spike detection above threshold |
| Per-DSP demand trends, identifies buyer pullbacks and timing |
| Composite health score combining all four signal dimensions |
Demo

Architecture
This is an MCP tool server - it exposes diagnostic capabilities (data retrieval and analysis) but does not orchestrate them. The LLM (Claude, GPT, or any MCP-compatible client) decides which tools to call, interprets results, and generates the diagnosis. Separating the intelligence layer from the data layer makes the server interoperable with any AAMP-compatible agent.
LLM (Claude) → MCP protocol → supply-ops-agent → mock dataSee also the IAB Tech Lab Seller Agent SDK for the reference supply-side agent implementation that this project complements.
Diagnostic logic is grounded in operational patterns from SSP exchange ops: bid stream anomalies, sellers.json reconciliation, and IVT taxonomy aligned with HUMAN/TAG standards.
Baked-in incidents
The mock data includes three realistic scenarios for testing:
Partner Bravo - Bid rate drops 40% on day 5, caused by undeclared seller IDs in ads.txt
Partner Delta - SIVT spikes to 18% on days 10-12, then returns to baseline
Partner Foxtrot - Two major DSP buyers pull back spend starting day 8
Setup
python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txtRunning
Standalone (stdio)
python server.pyWith Claude Desktop
Add to your Claude Desktop config (~/Library/Application Support/Claude/claude_desktop_config.json):
{
"mcpServers": {
"supply-ops-agent": {
"command": "python",
"args": ["/absolute/path/to/supply-ops-agent/server.py"]
}
}
}With Claude Code
claude mcp add supply-ops-agent python /absolute/path/to/supply-ops-agent/server.pyExample queries
Once connected, try asking:
"Check the health of Partner Bravo"
"Why did Bravo's bid rate drop?"
"Is there any IVT issue with Delta?"
"Which DSPs pulled back from Foxtrot?"
"Run a full health check on all partners"
Tech stack
Python 3.10+
MCP Python SDK (
mcp)Synthetic mock data (no external dependencies)
This server cannot be installed
Resources
Unclaimed servers have limited discoverability.
Looking for Admin?
If you are the server author, to access and configure the admin panel.
Latest Blog Posts
MCP directory API
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/mpessis/supply-ops-agent'
If you have feedback or need assistance with the MCP directory API, please join our Discord server