autoresearch

autoresearch is a skill for Claude Code from biggora/claude-plugins-registry. It costs 236 tokens per session (3,469 once invoked), scanned A, original, MIT.

An automated trial-and-measure loop for improving anything with a score, such as prompts, configuration, code speed, or written copy. A human defines what success means, and the agent repeatedly tests possible changes.

In plain words
What is it for?
Use it to run repeated experiments, tune settings or prompts, compare variants, and keep changes that improve a measurable result.
Why use it?
It removes much of the manual guesswork from optimization by comparing experiments against one agreed measurement.

Skill for Claude Code

Written for Claude Code: shipped in a Claude Code plugin.

Part of the project-delivery-orchestrator plugin — 34 skills, 1 command, 5 agents shipped together

Good fit Use it to run repeated experiments, tune settings or prompts, compare variants, and keep changes that improve a measurable result.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/biggora/claude-plugins-registry/autoresearch
Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

Any agent
npx skills add biggora/claude-plugins-registry --skill autoresearch
Clone the repo
git clone --depth 1 https://github.com/biggora/claude-plugins-registry

Made for: Claude Code.

Or install project-delivery-orchestrator, the plugin that ships this one along with the rest of its 34 skills, 1 command, 5 agents.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for autoresearch

README.md
[![agentmods](https://agentmods.dev/badge/skills/biggora/claude-plugins-registry/autoresearch/github.svg)](https://agentmods.dev/skills/biggora/claude-plugins-registry/autoresearch)
Your own site
<a href="https://agentmods.dev/skills/biggora/claude-plugins-registry/autoresearch"><img src="https://agentmods.dev/badge/skills/biggora/claude-plugins-registry/autoresearch/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for autoresearch

Your own site · 80×15
<a href="https://agentmods.dev/skills/biggora/claude-plugins-registry/autoresearch"><img src="https://agentmods.dev/badge/skills/biggora/claude-plugins-registry/autoresearch.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 236 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 3,469 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00236 $0.03469
Opus 5 $0.00118 $0.01734
Sonnet 5 $0.00047 $0.00694
Haiku 4.5 $0.00024 $0.00347

Measured 9d ago against content hash 39152c188909, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-09, from the pricing page.

Security

Grade A, and why

autoresearch scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 9d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

src/skills/autoresearch/SKILL.md · 386 lines

How it starts

The opening of the file, as written. The whole thing — 386 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Overview

The Agent-Optimizer pattern, originating from Andrej Karpathy's autoresearch project, asks a deceptively simple question: what if an AI agent ran its own experiments indefinitely?

The insight is that most optimization work follows the same shape regardless of domain — you have something to improve, a way to measure improvement, rules about what you can change, and an isolated place to make changes. The agent's job is to run that loop faster and more systematically than a human would.

This skill makes that pattern available for any measurable goal: not just neural network research, but prompt tuning, email copy optimization, config parameters, CSS performance scores, query speed, or any other domain where "better" maps to a number.

The key shift: the human defines success criteria upfront. The agent handles the trial-and-error.


The Four Pillars

Every optimization project needs exactly these four things. If any pillar is missing, establish it before the loop begins.

1. Scalar Metric — The Score

A single number that goes up (or down) when things improve. Single means single — if you have multiple metrics, pick the one that matters most, or combine them into a weighted score.

Why scalar? Because the agent needs an unambiguous signal. "Better" must mean "higher number" (or lower, consistently).

Examples:

  • Code performance: execution time in milliseconds (lower = better)
  • Prompt engineering: LLM judge score 0-100 for output quality
  • Email copy: click-through rate % from A/B test simulation
  • Config tuning: requests-per-second under load
  • CSS/UI: Lighthouse performance score 0-100
  • Text quality: Flesch readability score
  • Test suite: pass rate % or coverage %

2. Evaluator — The Test

An automated script that takes the current state of the Sandbox and outputs the metric. The evaluator is sacred — the agent never modifies it.

Why immutable? Because if the agent can change the test, it will eventually find a way to pass the test without actually improving anything. The evaluator is the ground truth.

Read the full file on GitHub · 386 lines

Files

What ships with it

3 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 9d ago First seen · 386 lines · 236 tokens per session scan A 39152c188909

Subscribe to this mod's changes

autoresearch is a skill published in the GitHub repository biggora/claude-plugins-registry (2 stars, last pushed 11d ago), licensed MIT. It adds 236 tokens to every session and 3,469 once invoked, about $0.0012 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Related

Other skills, from other repositories

systematic-debugging

Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.

obra/superpowers · 21 tokens

local-ai-agents

Build local-first AI agents that run entirely on a developer workstation with Microsoft Foundry Local and Qwen function-calling models. Covers Small Language Models (SLMs), the OpenAI-compatible local endpoint, sandboxed local tools, local RAG with Chroma, local MCP servers, hybrid cloud/local routing, and the…

microsoft/ai-agents-for-beginners · 200 tokens

next-cache-components-adoption

Turn on Cache Components in a Next.js app and resolve the blocking routes it surfaces. Use when the user wants to enable, adopt, or migrate to Cache Components, flip the cacheComponents flag, work through a flood of blocking-prerender / instant validation errors, run the cache-components-instant-false codemod, or…

vercel/next.js · 95 tokens

insight-error-page

Write or audit an insight-kind error page for the Next.js dev overlay. Use when creating a new errors/ .mdx page, auditing an existing one, or checking that a page matches the framework fix cards. Covers page structure, title alignment, FixCard cards with Copy prompt button, code snippets, terminology verification…

vercel/next.js · 83 tokens

next-cache-components-optimizer

Drive a Next.js route to instant navigation by setting up an agentic loop, under Cache Components / PPR, on initial load (hard navigation) and client-side navigation (soft navigation). Encode the goal as a failing @next/playwright instant() e2e and work it to green, one verified route at a time; the shipped test then…

vercel/next.js · 170 tokens

next-partial-prefetching-adoption

Turn on Partial Prefetching in a Next.js app and work through the insights it surfaces. Use when the user wants to enable or adopt Partial Prefetching, flip the partialPrefetching flag, opt routes in with export const prefetch = 'partial', audit Link prefetch={true} behavior, preserve existing prefetched UI with…

vercel/next.js · 103 tokens