Top AI Repos — open-source AI, indexed and scored
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
An autonomous AI agent pipeline for rigorous academic research, featuring strict DOI verification, multi-agent Scopus/OpenAlex/Semantic Scholar retrieval, and APA 7th .docx generation
| Date | Stars |
|---|---|
| 2026-07-31 | 290 |
| 2026-08-05 | 291 |
| 2026-08-06 | 291 |
Today
— stars today
This week
— stars this week
This month
— stars this month
Momentum
0.0
growth rate 0.00%/day
# Deep Research Agent **Deep Science Writer** is an industrial-grade, end-to-end scientific research pipeline and AI agent skill. Designed for the Hermes/ECC framework, it completely automates the literature review process: from background full-text database querying down to anti-hallucination verification, Nature/Science-level peer review, `.docx` compilation, and Obsidian/NotebookLM knowledge base ingestion. ## 🌟 About An autonomous AI agent pipeline for rigorous academic research, featuring strict DOI verification, multi-agent Scopus/OpenAlex/Semantic Scholar retrieval, and APA 7th `.docx` generation. ## 🌟 Key Features * **Precision Background Sourcing (Two-Stage Pipeline):** Subagents first fetch and screen a large pool of abstracts (e.g., 100+). For the highly relevant subset, the agent downloads and deeply reads the *full text* (Methodology, Results) to prevent abstract-induced hallucination. **FULL-TEXT verification of final claims is absolutely mandatory.** * **Strict Quality Control:** Explicitly targets Q1-Q2 journals. Marks Q3 when strictly necessary. **Bans all Q4 and MDPI publications.** * **Multi-Agent Discovery:** Spawns exactly three concurrent subagents to query Scopus, Exa, OpenAlex, and Semantic Scholar simultaneously for exhaustive mapping. * **Strict Compliance:** The agent is hard-coded to strictly follow every step in order. Skipping phases or taking shortcuts (like drafting without full-text verification) is strictly forbidden. * **Zero-Hallucination Guarantee (Phase 4.5):** Automatically runs live HTTP `requests` tests against every generated DOI to ensure 100% validity. Cross-references generated claims against raw full texts to prevent AI overstatement. * **Automated Peer Review:** Integrates the `remi` peer-review skill *(named in tribute to my academic advisor, Remi Chauvy)* to aggressively strip "AI fluff" (e.g., "delve", "tapestry") and enforce rigorous academic tone. * **English Output Only:** Enforces strict English language generation for all academic reports and drafts, regardless of conversational language. * **Hands-Free Output:** Programmatically builds a fully formatted Microsoft Word document (`.docx`) with APA 7th hanging indents and auto-generated data visualizations. * **Knowledge Management Loop:** Automatically saves research summaries to your local **Obsidian Vault** and explicitly uploads **every individually cited reference** as a separate source into **Google NotebookLM** for precise audio overviews and cross-referencing. --- ## 📋 Prerequisites To run this skill successfully, your host environment must be configured with the following dependencies and Model Context Protocol (MCP) servers: ### 1. System Requirements * **Hermes Agent** (or a compatible ECC/Claude Code runner). * **Node.js** (v18+ recommended) and `npx` for running MCP servers. * **Python 3.10+** (in your system PATH or agent's virtual environment). ### 2. Python Dependencies The agent relies on Python to generate documents and verify links. ```bash pip install python-docx PyMuPDF requests matplotlib seaborn pandas ``` ### 3. Required MCP Servers Add these to your `config.yaml` or `claude_desktop_config.json`: * **Scopus MCP (`scopus-mcp`)** * Required for premium literature retrieval. * **API Key:** Requires a free Elsevier Scopus API Key (`SCOPUS_API_KEY`). Apply at the [Elsevier Developer Portal](https://dev.elsevier.com/). * **NotebookLM MCP (`notebooklm-mcp-server`)** * Required for Phase 7 knowledge ingestion. * **Auth:** Must run `npx notebooklm-mcp-server auth` once in your terminal to authenticate your Google Account locally. * **Exa Search MCP** (Highly Recommended) * Used for neural search fallback and broad open-access discovery. * **GitHub MCP & Playwright MCP** (Optional but recommended for broader functionality). ### 4. Local Environment * **Obsidian:** The skill looks for `%OBSIDIAN_VAULT_PATH%\Hermes\` (fallback: `%USERPROFILE%\Documents\Obsidian Vault\Hermes\`). Update `SK
Excerpt of 8,084 characters
Read on GitHubWould you bet a product on this? Bounded 0–100 and slow moving.
matched fp:1461f280bb8ce23b, desc:ai agent, desc:multi-agent, desc:multi agent