dsh-literature

The most tedious part of tracking literature is not the one-time search but the recurring one: checking whether new papers landed under your topic, recalling which DOIs you already screened, and keeping the journal whitelist in the right state. dsh-literature turns that circular chore into a linear pipeline. You state a topic or a researcher; the plugin pulls fresh output from Crossref and arXiv, you grade the hits, they land in a global library, and the next cycle starts from local records instead of cold remote calls.

Three layers of capability. For retrieval, Crossref remote search handles broad queries while DOI lookups hit the local library first, sparing redundant upstream requests. For tracking, you can pin a research topic (optionally scoped by an ISSN whitelist) or a specific researcher (matched via ORCID disambiguation), set a recurring schedule, and each firing triggers a two-source sweep, a human grading pass, and a write into the global library. For context, a built-in SCI journal directory (title, ISSN, impact factor, CAS zone), a cross-session researcher profile store, and session-scoped deadline reminders round it out. The plugin explicitly does not fetch full text or translate; it hands you structured metadata and lets you read what matters.

This is the tool for researchers already living inside the DeepSeek Harness workflow who need to keep a steady eye on one or two sub-fields or a handful of peers: graduate students monitoring an assigned area, review writers maintaining a living bibliography, or small groups screening submissions against journal zones. Enabled exclusively through the Literature Tracking Assistant preset, it stays invisible to standard sessions, so the only cost is one extra chain in your pipeline and zero surface area in everything else.