Skills: Literature Trends

Literature Trends Skill

Trace how explanatory hypotheses evolve over time within a tagged literature corpus.

Purpose

The literature-trends skill implements an abductive argumentation approach to literature analysis: given a set of papers tagged with a keyword, identify the dominant explanatory hypotheses in successive time windows and trace their genealogy — which hypotheses replaced which, which were refined, and which were abandoned.

This is not a citation analysis tool. It is a tool for understanding how the explanatory landscape shifts in a field.

Prerequisites

• TypeDB running (make db-start)
• uv installed
• Papers ingested and tagged via the scientific-literature skill (the trend analysis operates on scilit-paper entities already in TypeDB)

Commands

uv run python .claude/skills/literature-trends/literature_trends.py <command> [args]

Command	What it does
create-thread	Create a new trend thread for a keyword or topic
record-hypothesis	Record the dominant explanatory hypothesis for a time window
record-genealogy	Link two hypotheses (replacement, refinement, abandonment)
show-thread	Display the full hypothesis genealogy for a thread

Typical Workflow

You: I have papers on CRISPR off-target effects tagged in my corpus.
     Create a trend thread for this topic.

You: For the 2017-2019 window, the dominant explanation was
     non-specific nuclease activity at mismatched sites.
     Record this as hypothesis H1.

You: For 2020-2022, the field shifted to RNA-guided scanning models.
     Record this as H2 and note it refined H1.

You: Show me the full genealogy for the CRISPR off-target thread.

Schema

• lt-thread — a named trend analysis thread (domain-thing)
• lt-hypothesis — an explanatory claim associated with a time window
• lt-genealogy — relation linking hypotheses (replacement / refinement / abandonment)

Relationship to Other Skills

Literature Trends depends on scientific-literature for the underlying paper corpus. Papers must be ingested and tagged before trend analysis can begin. The literature-trends skill queries scilit-paper entities and their keyword attributes to identify the relevant corpus for a thread.