# Eclipse Sandclock

> Eclipse Sandclock is an open-source MCP server for agent tooling. Local stdio MCP that tracks when agent eval results expire, using fixture scorecards. It is available on CLI and self-hosted.

- **Kind:** MCP server
- **Pricing:** Open source
- **Platforms:** CLI, self-hosted
- **Website:** https://eclipsebazaar.com
- **Repository:** https://github.com/cyan-eclipse-technologies-llc/eclipse-sandclock-mcp
- **MCP Registry name:** io.github.CYAN-ECLIPSE-TECHNOLOGIES-LLC/eclipse-sandclock-mcp
- **Categories:** [agent tooling](https://ai-powered.com/categories/agent-tooling), [agent benchmarks](https://ai-powered.com/categories/agent-benchmarks)
- **Listing updated:** 2026-11-18

## About

Eclipse Sandclock works on the idea that a benchmark result goes stale. It reads Benchkit scorecards and drift evidence and reports when an evaluation should no longer be relied on, so an agent can tell a fresh pass from an expired one.

The registry entry says it runs only as a local stdio MCP server, and that both the scorecards and the drift evidence are fixtures. It is therefore a sandboxed demonstration of eval-expiry tracking and not a monitor of live production tools. No tool list is published.

It comes from Cyan Eclipse Technologies, whose Bazaar treats a verdict as something with a lifecycle: re-scan on version change, probe for drift, re-bench when a tool drifts. Sandclock appears to be a small piece of that approach, aimed at developers building trust or governance workflows for agent tools.

## Features

- **Eval expiry clock:** Calculates when a benchmark verdict should be treated as expired.
- **Benchkit scorecard input:** Works over Benchkit scorecards, currently supplied as fixtures.
- **Drift evidence:** Combines fixture drift evidence with scorecards to judge whether a verdict still holds.
- **Local stdio transport:** Runs as a local stdio MCP server, with no hosted endpoint.

## Good for

- Prototyping verdict-expiry logic for agent tool certification
- Testing how an agent reacts to a stale benchmark result
- Experimenting with drift-aware trust checks on fixture data
- Learning how continuous trust operations might work for MCP tools

## FAQ

### How does Eclipse Sandclock run?

As a local stdio MCP server only. The registry entry lists no remote endpoint or published package.

### Does it check live tools?

No. It operates over fixture Benchkit scorecards and fixture drift evidence.

### Who makes it?

Cyan Eclipse Technologies LLC, which also runs the Cyan Eclipse Bazaar marketplace.

### Is Eclipse Sandclock free?

Yes. Eclipse Sandclock is open source, so you can use and self-host it for free.

### What are the best alternatives to Eclipse Sandclock?

The closest alternatives to Eclipse Sandclock are Eclipse Epitaph, Eclipse Benchkit, Eclipse Ward, Eclipse Greenwash and Eclipse Shelfprice.

## Where this comes from

- Description: written from the project's website and its MCP Registry entry.

## Alternatives

- [Eclipse Epitaph](https://ai-powered.com/mcp/eclipse-epitaph): Eclipse Epitaph is an MCP server for agent tooling.
- [Eclipse Benchkit](https://ai-powered.com/mcp/eclipse-benchkit): Eclipse Benchkit is a free MCP server for developer tools.
- [Eclipse Ward](https://ai-powered.com/mcp/eclipse-ward): Eclipse Ward is an MCP server for agent tooling.
- [Eclipse Greenwash](https://ai-powered.com/mcp/eclipse-greenwash): Eclipse Greenwash is an MCP server for climate and sustainability.
- [Eclipse Shelfprice](https://ai-powered.com/mcp/eclipse-shelfprice): Eclipse Shelfprice is an MCP server for e-commerce.
- [Eclipse Timecapsule](https://ai-powered.com/mcp/eclipse-timecapsule): Eclipse Timecapsule is an MCP server for agent tooling.
- [Eclipse Holdfast](https://ai-powered.com/mcp/eclipse-holdfast): Eclipse Holdfast is an open-source MCP server for finance and payments.
- [Eclipse Scapegoat](https://ai-powered.com/mcp/eclipse-scapegoat): Eclipse Scapegoat is an MCP server for agent tooling.

---
Source: https://ai-powered.com/mcp/eclipse-sandclock
