Autoresearcher — Research agent for Claude Code
One-bounded-dig research experimenter — reads the brief, proposes ONE hypothesis targeting ONE harness component, fetches ≤5 sources, edits ONE section, returns a scored diff.
How to install Autoresearcher
Installs to ~/.claude/agents/frankxai-frankx.ai-vercel-website-autoresearcher.md
mkdir -p ~/.claude/agents && curl -fsSL https://raw.githubusercontent.com/frankxai/frankx.ai-vercel-website/HEAD/.claude/agents/autoresearcher.md -o ~/.claude/agents/frankxai-frankx.ai-vercel-website-autoresearcher.md Restart Claude Code, or start a new session, for it to be picked up.
What Autoresearcher does
name: autoresearcher description: One-bounded-dig research experimenter — reads the brief, proposes ONE hypothesis targeting ONE harness component, fetches ≤5 sources, edits ONE section, returns a scored diff. Karpathy-autoresearch pattern. tools: Read, Write, Edit, Grep, Glob, WebSearch, WebFetch, Bash model: sonnet
You are an **autoresearch experimenter** for the FrankX research hub. Your job is to propose ONE small improvement to a research brief, measure it, and return the result. Y
Alternatives in Research
- Worker Researcher — Web research agent 1k ★
- Swarm Researcher — READ-ONLY research agent - discovers tools, fetches docs, stores findings 508 ★
- Deep Research Fetcher — Fetches one disjoint batch of URLs into a nav-deep-research run as fenced source notes 232 ★
Full documentation available on GitHub
View Source RepositoryRelated Agents
Harness Optimizer Investigator
Investigate one selected harness-optimizer evidence lane within a bounded read-only packet.
Errand
Bounded read-only lookup — docs, MCP/skill queries, tool runs, research. Answers, never edits.
Coa Analysis
Generate and compare courses of action. Human-reviewed, never targeting.
Klein Experimenter
EXPERIMENT-loop worker for Klein Auto Research — runs disciplined one-candidate v2 transactions with immutable
Klein Referee
REFEREE worker for Klein Auto Research — Gate 3. Reviews a synthesized study in a FRESH context on a different
Prior Scores Finder
Finds prior scored reforms to anchor a new analysis — PolicyEngine's own published scores first, then official
Related Skills
Autoresearch
Claude Autoresearch Skill — Autonomous goal-directed iteration for Claude Code. Inspired by Karpathy's autores
Autovoiceevals
A self-improving loop for voice AI agents. Uses karpathy's autoresearch as foundation.
Autoresearch Anything
Claude Code skill: autonomous experimentation pipeline for any business metric using Karpathy's autoresearch p