Use this skill when the user asks to "research background on", "find background information about", "search web for context on", "look up policy history of", "find grey literature on", "research institutional context", "find government reports on", "background research for introduction section", "research country context", "web research on", "autonomous web search", "deep research on", "find sources about", or needs web-sourced background context for academic writing (political economy, comparative politics, economic history, public policy). Complements Zotero/arXiv for web-accessible grey literature and policy documents.
How to install gpt-researcher?
haipingxu/social-science-claude-scholar/gpt-researcher
$npx -y skills add haipingxu/social-science-claude-scholar --skill gpt-researcherInstalls into the current project.
›Prefer a prompt? Paste this to your agent
Run `npx skills use "https://github.com/haipingxu/social-science-claude-scholar" --skill "haipingxu/social-science-claude-scholar/gpt-researcher"` and follow the generated skill instructions now. Read its complete output, redirecting it to a temporary file first if necessary. Resolve relative paths from the supporting-files directory it provides.
Use the skills in "https://github.com/haipingxu/social-science-claude-scholar" that are relevant to the current task. Run `npx skills add "https://github.com/haipingxu/social-science-claude-scholar"` and select the relevant skills, then follow their instructions.
SKILL.md
| 1 | # GPT Researcher |
| 2 | |
| 3 | GPT Researcher is an autonomous agent designed for comprehensive online research on any given topic. |
| 4 | |
| 5 | The agent produces detailed, factual and unbiased research reports with citations, aggregating over 20 web sources per task to form objective conclusions. |
| 6 | |
| 7 | ## Why |
| 8 | |
| 9 | - Manual research takes hours; GPT Researcher runs in ~3 minutes. |
| 10 | - Single-source answers hallucinate; aggregation across 20+ sources reduces bias. |
| 11 | - Reports export straight to Markdown, PDF and Docx. |
| 12 | |
| 13 | ## Architecture |
| 14 | |
| 15 | A planner agent generates research questions. Execution agents crawl and summarize per-question. The planner aggregates everything into one filtered, cited report. |
| 16 | |
| 17 | ## Quickstart |
| 18 | |
| 19 | ```bash |
| 20 | pip install gpt-researcher |
| 21 | export OPENAI_API_KEY=... |
| 22 | gpt-researcher --query "should I invest in open-source AI infra?" |
| 23 | ``` |
Deep-research mode: recursive sub-queries with citation dedup.
2026-07-19
MCP server mode; Claude and Gemini provider fixes.
2026-07-02
Docx export, source-quality scoring, 2x faster crawls.
2026-06-11
