interrank

Query AgMoDB snapshot data for model and benchmark rankings.

3|1|Updated Feb 15, 2026

Search Tags:#models #evaluation #benchmarks #snapshot #ranking #AgMoDB

Ready to run

npx skills add https://github.com/mistakeknot/Demarch --skill interrank

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: interrank
Source: https://github.com/mistakeknot/Demarch/tree/main/.gemini/generated-skills/interrank
Command: npx skills add https://github.com/mistakeknot/Demarch --skill interrank

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill provides a snapshot-backed ranking mechanism for models and benchmarks, specifically designed for the AgMoDB system, ensuring efficient and consistent evaluation.

Core Features & Use Cases

Snapshot-backed Ranking: Queries AgMoDB snapshot data for model and benchmark rankings.
Read-only MCP Plugin: Designed for querying existing data without direct modification.
Use Case: Quickly retrieve the performance ranking of different AI models on a specific benchmark from a historical snapshot to inform future development decisions.

Quick Start

Use the interrank skill to get the ranking for the 'bert-base' model on the 'mnli' benchmark.

Frequently Asked Questions about interrank

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I retrieve historical model rankings from AgMoDB snapshots?▼

To retrieve historical model rankings from AgMoDB snapshots, you query the snapshot-backed data specifically for model and benchmark performance metrics. This read-only approach ensures consistent evaluation without modifying the underlying database.

What is the best way to compare AI model performance across benchmarks?▼

Comparing AI model performance across benchmarks is best handled by querying historical snapshot data. This mechanism provides a consistent, read-only ranking of models to inform future development decisions based on past evaluations.

Can I use a read-only MCP plugin to query existing benchmark data?▼

Yes, you can use a read-only MCP plugin to query existing benchmark data. It is specifically designed to retrieve model and benchmark rankings from AgMoDB snapshots without allowing direct modification of the stored information.

How do I handle unknown metric or benchmark keys when evaluating model rankings?▼

Handling unknown metric or benchmark keys when evaluating model rankings requires explicit management within your query. The system mandates strict adherence to snapshot compatibility, ensuring unknown keys do not disrupt the evaluation process.

How do I get the performance ranking for a specific model like bert-base on the mnli benchmark?▼

To get the performance ranking for bert-base on the mnli benchmark, you execute a targeted query against the AgMoDB snapshot data. This retrieves the specific historical ranking results for that model and benchmark combination.