Course Drift Oracle MCP server
{
"mcpServers": {
"710git-course-drift-oracle": {
"type": "http",
"url": "https://signetworks.atelieri.workers.dev/mcp"
}
}
}{
"mcpServers": {
"710git-course-drift-oracle": {
"type": "http",
"url": "https://signetworks.atelieri.workers.dev/mcp"
}
}
}{
"mcpServers": {
"710git-course-drift-oracle": {
"type": "http",
"url": "https://signetworks.atelieri.workers.dev/mcp"
}
}
}{
"servers": {
"710git-course-drift-oracle": {
"type": "http",
"url": "https://signetworks.atelieri.workers.dev/mcp"
}
}
}{
"mcpServers": {
"710git-course-drift-oracle": {
"type": "http",
"url": "https://signetworks.atelieri.workers.dev/mcp"
}
}
}file .mcp.json (in your project root)file claude_desktop_config.jsonfile ~/.cursor/mcp.jsonfile .vscode/mcp.jsonfile ~/.codeium/windsurf/mcp_config.jsonversion 1.0.0
What it does
Course Drift Oracle scans Microsoft's ai-agents-for-beginners course for deprecated model pins and package version mismatches that don't appear in diffs. It provides signed drift reports (free summary; $0.25 for full findings) and audits public websites for AI-agent readiness using 12 checks.
- Check whether a course's pinned models are deprecated or scheduled to retire before they break in production.
- Verify that package version requirements in course materials still resolve to the versions the lessons were written against.
- Audit a public website for AI-readiness by checking robots.txt directives, llm.txt format, structured data, and other agent-discovery signals.
Tools this server exposes
| Tool | Does |
|---|---|
| drift_summary | Free tool that returns finding counts by severity and lesson, plus a signed receipt. |
| verify_report | Free tool that verifies a receipt's Ed25519 signature and optionally checks that a findings array matches the signed hash. |
| drift_report | Returns every drift finding with file, line number, diagnosis, and fix; paid $0.25 via MPP testnet. |
| drift_report_x402 | Same as drift_report but paid via x402/USDC on Base Sepolia testnet. |
| model_status | Free tool that returns a signed catalog entry for a model ID showing current, deprecated, retired, or unknown status. |
| model_status_batch | Batch lookup of multiple model IDs with per-model status and an action-needed gate boolean; paid $0.10 via MPP testnet. |
| model_status_batch_x402 | Same as model_status_batch but paid via x402/USDC on Base Sepolia testnet. |
| site_audit_summary | Free tool that returns an agent-readiness score for a public site (counts and target date only; detail is the paid tier). |
| site_audit | Full audit of all 12 website checks with per-check status, detail, and fix list as a signed receipt; paid $0.25 via MPP testnet. |
| site_audit_x402 | Same as site_audit but paid via x402/USDC on Base Sepolia testnet. |
Before you enable it
An MCP server runs with whatever access you give it. Start with a development account rather than production, and check which tools the server exposes before you let an agent call them unattended.