Generates a chronological research development-history article from database retrieval and web search.
npx skills add https://github.com/zjunlp/Mechanist --skill mhistory-zjunlp
Please help me install this Agent Skill. Skill: mhistory Source: https://github.com/zjunlp/Mechanist/tree/main/skills/mhistory Command: npx skills add https://github.com/zjunlp/Mechanist --skill mhistory-zjunlp
High-intent search queries and answers about installing and using this skill.
Invoke /mhistory followed by your topic in English or Chinese. The skill runs two parallel database retrieval passes plus WebSearch, then writes a 2500β4500 word markdown article to development_history.md.
It queries the mechanic-db SEARCH service twice in parallel: once with temporal_mode=history for even coverage across eras, and once with temporal_mode=recent for recency-boosted frontier work. WebSearch fills gaps for pre-database classics and the last 1β6 months of arXiv papers.
Both database passes degrade gracefully by writing a skipped flag to their output files. The article is then built from WebSearch results and the model's own knowledge, and the final markdown explicitly notes that mechanic-db was unavailable.
The markdown article is organized by chronological eras or thematic phases, includes a TL;DR, a contemporary landscape section, a frontier section tagged by source, a Tensions and Open Questions section, and a references list with DOI or arXiv URLs.
Yes, the topic input is free text in Chinese or English and the pipeline is domain-agnostic. However, retrieval quality depends on the coverage of the mechanic-db corpus, which is oriented toward research literature.
SebastianElvisAsk targeted questions to clarify research goals from papers or topics.
LittleSongxxGenerate structured research proposals and grant applications from literature review outputs.
acertainKnightSynthesize research papers into structured literature reviews with citation-context analysis.
kbaseincubatorSearch and synthesize biological literature across PubMed, arXiv, bioRxiv, and Google Scholar.
