ar-start

Initialize autoresearch loops for prepared benchmark projects.

Updated Aug 27, 2026
One-click install
npx skills add https://github.com/Elijah-J/InfoScraper --skill ar-start
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: ar-start
Source: https://github.com/Elijah-J/InfoScraper/tree/main/.agents/skills/ar-start
Command: npx skills add https://github.com/Elijah-J/InfoScraper --skill ar-start

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Autoresearch workflows for new benchmark projects require coordinated setup, branch management, and initial evaluations. ar-start automates starting the loop from a prepared scaffold, reducing manual setup time and ensuring repeatable baselines.

Core Features & Use Cases

  • Bootstrap-and-run workflow: switch to the expected autoresearch/<slug> branch, optionally run bootstrap, and prepare the project for evaluation.
  • Baseline and first eval: create or update a baseline results row and execute the initial evaluation.
  • Guardrails and reporting: enforce workflow guardrails, report baseline scores, guardrail findings, and update project status.

Quick Start

Start the autoresearch workflow for a prepared project directory by running ar-start <project-dir> to initialize and launch the loop.

Frequently Asked Questions about ar-start

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

FAQPage Schema
How do I bootstrap an autoresearch loop for a new benchmark project?

To bootstrap an autoresearch loop, run ar-start on your prepared project directory to switch to the autoresearch/<slug> branch, execute bootstrap, and generate baseline evaluation results.

What files do I need to set up before starting an autoresearch workflow?

Starting an autoresearch workflow requires locating autoresearch.toml, program.md, and eval.py in your prepared project directory to ensure proper workflow orchestration and initial evaluation execution.

How does baseline generation work in benchmark optimization workflows?

Baseline generation in benchmark optimization creates or updates a baseline results row, executes the first evaluation, and reports baseline scores alongside guardrail status to establish repeatable performance metrics.

Can I use automated workflow orchestration to manage git branches for evaluation runs?

Yes, automated workflow orchestration manages git branches by switching to the expected autoresearch/<slug> branch before running bootstrap and executing initial evaluations within a dedicated workspace.

What is the best way to automate initial evaluation setup for benchmark projects?

The best way to automate initial evaluation setup is using a bootstrap-and-run workflow that enforces guardrails, executes the first eval, and reports baseline scores to ensure repeatable results.

Why does my autoresearch workflow require guardrails during baseline generation?

Autoresearch workflows require guardrails during baseline generation to enforce workflow requirements, verify guardrail findings, and update project status accurately after executing the initial evaluation.