doubao-academic-researcher

Conducts verified multi-stage academic literature research with citation validation and structured synthesis.

2|Updated Aug 9, 2026
One-click install
npx skills add https://github.com/DeepJH/doubao-skill-and-info --skill doubao-academic-researcher-deepjh
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: doubao-academic-researcher
Source: https://github.com/DeepJH/doubao-skill-and-info/tree/main/skills/doubao-academic-researcher
Command: npx skills add https://github.com/DeepJH/doubao-skill-and-info --skill doubao-academic-researcher-deepjh

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve? It helps researchers, students, and paper writers who have not yet locked in a specific thesis topic systematically investigate an academic direction, producing conclusion-first, citation-verified research surveys instead of unreliable or hallucinated literature summaries. ## Core Features & Use Cases - Verified Literature Retrieval: Searches across multiple perspectives, verifies every citation for existence and accuracy, and rejects unverifiable or hallucinated references. - Four-Stage Gated Pipeline: Orchestrates literature scouting, evidence synthesis, review drafting, and Feishu (Lark) document delivery through script-enforced workflow gates and handoff validation. - Structured Deliverables: Produces core conclusions, a thematic literature map table, a Mermaid research-lineage logic graph, controversy and gap analysis, research direction suggestions, and a 3-6 paragraph review draft. - Use Case: A graduate student asks to survey the last three years of RAG applications in smart education with at least five typical case studies; the Skill decomposes the requirements, retrieves and verifies sources, synthesizes themes, and delivers a Feishu document with a logic graph and reference list. ## Quick Start Ask the assistant to research the current state, controversies, and open directions of a chosen academic topic and deliver a verified literature survey as a Feishu document.

Frequently Asked Questions about doubao-academic-researcher

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I run a verified academic literature survey with this workflow?

Start by initializing the workflow with a research brief, then decompose the request into a requirement checklist JSON. The pipeline runs literature scouting, synthesis, review writing, and document delivery stages, each gated by workflow.py validation scripts.

How does the skill prevent hallucinated citations in literature reviews?

Every citation passes a three-step verification: existence search by title and first author, accuracy check against the claim, and a status label. Unverifiable or gray-zone references are excluded, and DOI mismatches are treated as hallucination signals.

Can this skill write my thesis or literature review chapter for me?

No. It only delivers structured research results plus a 3-6 paragraph review draft as a format example for a broad direction. It refuses to produce full IMRaD papers, ghostwrite thesis paragraphs, or write the literature review chapter of your specific paper.

What output formats does the academic research workflow produce?

It produces a structured report with core conclusions, a literature map table, a Mermaid whiteboard logic graph, controversy and gap analysis, research directions, a continuous review draft, and by default a Feishu (Lark) cloud document that is read back and self-checked.

Why does the workflow get blocked before the literature search stage?

The literature-scout stage requires a valid requirement_checklist.json with at least one main-priority requirement and legal priority and in_scope values. Missing or malformed checklists return BLOCKED: NEED_REQUIREMENT_CHECKLIST until Step 0 decomposition is completed.

How is evidence strength graded across different academic disciplines?

Evidence is graded on two axes: a seven-level research-design pyramid and within-discipline fit against that field's gold standard. This prevents downgrading humanities primary sources or qualitative studies merely because they are not randomized controlled trials.