Tag
ScholarCatalyst introduces a benchmark where 184 lead authors label which papers advanced their completed CS projects, and shows that agentic search and even Claude Fable 5.1-based agents underperform simple embedding retrieval (0.42/0.51 vs 0.48 Recall@20) at retrieving idea-inspiring prior work.
TraceML is a tool for trajectory-level analysis of ML research agents, comparing Codex agent runs with Kaggle Grandmaster notebooks and finding that human experts explore broader strategy sets while agents collapse into narrow loops. Distilling human research skills into a ~1,000-token planning prompt lifted Codex scores on five of seven competitions; the work will be presented at NeurIPS 2026.
jevgrep is a research agent CLI powered by jev from typesafeai that reduces coding agent costs by 40% as verified on SWE-bench, with recent support for opencode zen added via a community PR.
Hyperresearch is an open-source Python tool that acts as a deep research harness for Claude Code and OpenAI Codex, using agents to collect, search, and synthesize web research into a persistent, searchable wiki with adversarial auditing.
Introducing jevgrep, a research agent CLI powered by jev that reduces coding agent costs by 40% as verified on SWE-bench.
jevgrep is a research agent CLI powered by jev from typesafeai that reduces coding agent costs by 40% by providing code context through natural language questions.
A research agent used for deal screening confused two startups named 'Blue' due to ambiguous naming, leading to mixed data; a simple fix was to skip internet search when trusted founder data is available.
This paper introduces Research Attention Prediction (RAP), a benchmark for evaluating LLMs in predicting future research attention shifts, revealing biases in evidence acquisition strategies.
This paper introduces Scientific Agent Skills, an open-source library of 163 procedural knowledge skills across 16 scientific domains to help AI research agents perform defensible analyses.
This article recommends paying attention to AI Scientist and discusses how research agents can learn from failures by analogizing to fuzz testing, thereby mapping the unknown and guiding subsequent experiments.
A Web Search Agent on nimble_search automates market research by scanning live public sources to provide structured snapshots including competitors, pricing, user complaints, and market gaps, saving builders from manual tab-hopping.
Matthew Stublefield of Fieldway built a product research agent using Tavily's web search API integrated with Claude Code via MCP, reducing a 5-hour manual cycle to 25 minutes and cutting weeks of work to 10 hours.
An open-source full-stack research agent example using Gemini, LangGraph, React, and Google Search, demonstrating an end-to-end workflow with dynamic query generation, reflection loops, and cited answers.
Open Deep Research is an open-source Next.js app that combines Firecrawl Search and Extract with a reasoning model to perform iterative web research loops, including search, extraction, planning, source tracking, and synthesis.
This paper investigates what makes a small 4B parameter model's citations faithful when used as an on-device research agent, finding that exposure to more source text improves faithfulness while retrieval recall limits coverage.
开源的科研智能体 OpenAI4S，通过持久化内核运行真正的 Python/R 代码，结合 JSON 工具编排，让研究者可以用极低成本（豆包 ¥9.9/月）复现 Claude Science 级别的科研自动化。
This paper introduces ResearchStudio-Idea, a reusable research ideation skill suite based on analysis of ML conference papers. It includes Paper-Search, Scoop-Check, and IdeaSpark, generating evidence-supported research ideas using 15 patterns extracted from 1,947 papers.
LangChain demonstrates an agent that drafts a cited VC investment memo in ~90 seconds for $0.40 using the Perplexity Agent API and LangGraph, with four parallel research nodes and a synthesizer.
Built a research agent from scratch without frameworks, available on GitHub.
A developer reimplements Arbor, a multi-agent research system from a paper, on LangGraph, sharing lessons about graph structure enforcing invariants and backpropagating insights from failures. The reimplementation, arbor-lg, uses LangGraph's StateGraph and SQLite checkpointer for durability.