comprehensive-survey

name	comprehensive-survey
description	Automated full-spectrum research survey. Uses litian-academic-search (7-source) for academic papers + WebSearch site: filters for social media. Ingests all found papers via paper-read --light. Produces structured FINAL_REPORT.md per topic + concept glossary with pseudocode. Use when user wants "全面调研", "综合调研", "survey this topic", "调研一下", "帮我全面了解一下这个方向".

Comprehensive Survey Pipeline

Automated multi-source survey: academic papers via litian-academic-search + social media via WebSearch + auto-ingestion into OmniBox.

Pipeline

Topic(s) → litian-academic-search (all sources) + social WebSearch
  │
  ├─ P1: Academic search per topic
  ├─ P2: Social platform search (WebSearch site: filters)
  ├─ P3: Ingest all found papers → paper-read --light
  ├─ P4: Comprehensive paper list (dedup, sort, link)
  ├─ P5: Concept extraction + glossary with pseudocode
  ├─ P6: Compile FINAL_REPORT.md per topic
  └─ P7: README + verification

Parameters

Parse from $ARGUMENTS:

• Topic(s): one or more research directions
• --depth <quick|standard|deep> — papers per topic: quick=15, standard=30, deep=50. Default: standard.
• --output <dir> — output directory

Phases

P1: Academic Search

For each topic:

/litian-academic-search "<TOPIC>" --sources all --k <K>
/litian-academic-search "<topic> survey review" --sources arxiv,s2,web --k 5

K = depth: quick=8, standard=15, deep=25.

P2: Social Search

WebSearch "<中文关键词> site:bilibili.com"
WebSearch "<topic> site:zhihu.com"
WebSearch "<english topic> site:x.com"
WebSearch "<topic> site:reddit.com"
WebSearch "<topic> blog deep dive"

P3: OmniBox Ingestion

For every top-k paper found: paper-read <arxiv_id> --topic <detected> --light

P4: Comprehensive Paper List

Deduplicate across all topics + sources. Write paper_list_comprehensive.md.

P5-P7: Concept Extraction + Reports + README

Standard structure. See references/output-spec.md for exact formats.

Depth Presets

Preset	Papers/topic	litian-academic-search --k	Social	Report
quick	~15	8	Skip or 1-2	500+ lines
standard	~30	15	All 5 platforms	800+ lines
deep	~50	25	All 5 + deep dive	1000+ lines

Calls

• litian-academic-search (P1)
• paper-read --light (P3)

Called by

manual invocation, survey-to-deck (as step ① of the pipeline)