autoresearch

autoresearch is a skill for Claude Code, Codex from pbc-os/smb-starter-kit. It costs 63 tokens per session (5,489 once invoked), scanned B, original, MIT.

An automated experiment loop for improving a system when you can measure its results. Separate agents generate changes, check them against pass/fail rules and held-out tests, and review how the search is going.

In plain words
What is it for?
Use it to tune parameters, compare configurations, explore improvements, and continue searching when a single best result is not enough.
Why use it?
It reduces the need to run experiments and judge results by hand, while helping avoid overfitting to the tests used during tuning.

Skill for Claude CodeCodex

Written for no agent in particular: nothing here depends on one.

Good fit Use it to tune parameters, compare configurations, explore improvements, and continue searching when a single best result is not enough.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/pbc-os/smb-starter-kit/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 pbc-os/smb-starter-kit --skill autoresearch
Clone the repo
git clone --depth 1 https://github.com/pbc-os/smb-starter-kit

Made for: Claude Code, Codex.

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/pbc-os/smb-starter-kit/autoresearch/github.svg)](https://agentmods.dev/skills/pbc-os/smb-starter-kit/autoresearch)
Your own site
<a href="https://agentmods.dev/skills/pbc-os/smb-starter-kit/autoresearch"><img src="https://agentmods.dev/badge/skills/pbc-os/smb-starter-kit/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/pbc-os/smb-starter-kit/autoresearch"><img src="https://agentmods.dev/badge/skills/pbc-os/smb-starter-kit/autoresearch.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 63 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 5,489 The whole file, excluding the scripts and references it only reads on demand.
Security scan B 1 finding. 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 $0.00063 $0.05489
Opus 5 $0.00032 $0.02745
Sonnet 5 $0.00013 $0.01098
Haiku 4.5 $0.00006 $0.00549

Measured 10d ago against content hash 7b84a634e731, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-10, from the pricing page.

Security

Grade B, and why

autoresearch scanned grade B with 1 finding 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 10d 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.

Asks for rootmediumPrivilege escalation

A mod that escalates privileges can change anything on the machine, not only the project.

chmod 444 eval.py # or eval.sh, eval.js, etc.
skills/tier-5-automation/autoresearch/SKILL.md · 414 lines

How it starts

The opening of the file, as written. The whole thing — 414 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Autoresearch

Autonomous iterative improvement for any measurable system.

Built on Karpathy's autoresearch pattern, extended with:

  • Critic agent with binary pass/fail gates and holdout validation — generator/verifier separation so the agent can't grade its own homework
  • Coverage-driven exploration instead of pure hill-climbing — don't spend 40 experiments tuning one knob when another knob has never been touched
  • Metacognitive self-modification — a meta-reviewer agent that can improve the researcher's own optimization strategy between runs
  • Archive of stepping stones — maintain a diverse population of good configurations, not just the single best, so you can branch out when stuck in a local optimum

See the references/ folder for the academic and practitioner sources behind each of these extensions.

Triggers

  • "autoresearch"
  • "run autoresearch on..."
  • "iteratively improve..."
  • "optimize [X] autonomously"
  • "run experiments on..."
  • "self-improve [system]"
  • "find better parameters for..."
  • "tune [X]"
  • User wants to autonomously improve any measurable system

The Pattern

Three agents work in a pipeline, coordinated by a small set of files on disk:

┌─────────────────────────────────────────────────────────────────┐
│                                                                 │
│  ┌─────────────┐    ┌─────────────┐    ┌──────────────────┐     │
│  │  RESEARCHER │───▶│   CRITIC    │───▶│  META-REVIEWER   │     │
│  │  (proposes) │    │ (validates) │    │ (every N exps)   │     │
│  └─────────────┘    └─────────────┘    └──────────────────┘     │
│        │                   │                    │              │
│        ▼                   ▼                    ▼              │
│  1. Analyze weakness  4. Holdout gate      7. Review full log  │
│  2. Form hypothesis   5. Overfitting check 8. Update strategy  │
│  3. Change 1 param    6. Binary pass/fail  9. Suggest new      │
│     + run eval           → KEEP/DISCARD       dimensions       │
│        │                   │                    │              │
│        └───────────────────┴────────────────────┘              │
│                            │                                   │
│                    ┌───────▼────────┐                          │
│                    │ COVERAGE MAP   │                          │
│                    │ What's been    │                          │
│                    │ explored?      │                          │
│                    │ What's missing?│                          │
│                    └────────────────┘                          │
│                            │                                   │
│                    ┌───────▼────────┐                          │
│                    │    ARCHIVE     │                          │
│                    │ Top N configs  │                          │
│                    │ Branch from    │                          │
│                    │ diverse parents│                          │
│                    └────────────────┘                          │
└─────────────────────────────────────────────────────────────────┘

Read the full file on GitHub · 414 lines

Files

What ships with it

6 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. 10d ago First seen · 414 lines · 63 tokens per session scan B 7b84a634e731

Subscribe to this mod's changes

autoresearch is a skill published in the GitHub repository pbc-os/smb-starter-kit (10 stars, last pushed 3mo ago), licensed MIT. It adds 63 tokens to every session and 5,489 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it B with 1 finding (asks for root). 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