run-query

Query run-index.yaml and vault claims for experiment metadata by tag, date, type, or claim.

39|4|Updated Feb 3, 2026
One-click install
npx skills add https://github.com/swarm-ai-safety/swarm --skill run-query
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: run-query
Source: https://github.com/swarm-ai-safety/swarm/tree/main/.skills/run-query
Command: npx skills add https://github.com/swarm-ai-safety/swarm --skill run-query

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill allows users to efficiently search and retrieve information about past experiments and their associated data stored in a vault, without needing to load the raw data itself.

Core Features & Use Cases

  • Flexible Searching: Supports queries by tags, dates, experiment types, or specific claims.
  • Metadata Retrieval: Returns essential run metadata, including pointers to raw data, for quick analysis.
  • Use Case: Find all experiments tagged 'deception' conducted in March 2024 to review their parameters and key findings.

Quick Start

Query the run index for runs tagged 'calibration'.

Frequently Asked Questions about run-query

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

FAQPage Schema
How do I search past experiment history without loading raw data?

Search past experiment history by querying the run index and vault to retrieve run metadata and pointers. This approach returns essential information without loading raw data, facilitating efficient review of past multi-agent system experiments.

What's the best way to find experiments tagged with specific claims in a vault?

Find experiments tagged with specific claims in a vault by using free-text search across run-index.yaml and vault claims. This supports comprehensive historical analysis by matching tags, dates, experiment types, or specific claims.

Can I filter experiment runs by date and type in the run index?

Filter experiment runs by date and type in the run index using the flexible querying system. It supports searches by tags, dates, experiment types, or specific claims to retrieve targeted run metadata efficiently.

How does querying run metadata compare to loading full experiment vaults?

Querying run metadata returns essential run metadata and pointers instead of loading full experiment vaults. This distinction enables quick analysis of past multi-agent system experiments while avoiding the overhead of processing raw data.

Why does my free-text search across run-index.yaml return no results?

Free-text search across run-index.yaml may return no results if the tags, dates, types, or claims do not match existing experiment entries. Ensure the vault contains properly indexed run metadata matching your query parameters.