study

name

study

description

Generate a paper dossier. Study one paper (or idea with a seed paper) into a fixed-template HTML page in the anu atlas (~/.anu/atlas/papers/<id>): what it claims, what it stands on and led to, the nearest work classified by closeness, code that implements it, and the bounded gap. Fans out reader/lineage/neighborhood/code agents via the Workflow tool and screens neighbors cold. Use when the user runs /study or asks to catch me up on / understand / situate a paper. The /map of an idea.

/study: paper dossier

Produce a dossier of one paper and render it as one HTML page in the anu atlas, the literature twin of /map. One fixed template for every paper, always the same visual grammar: the paper (claim, method, result), nearest work by closeness, lineage (stands on · led to), implementations, and the gap. The page is the artifact; keep the chat output minimal.

Follow the finding-prior-art discipline throughout. Never name a paper or claim what it says from memory. Every record comes from the scholar engine (~/.local/share/anu/plugins/ideas/bin/scholar) or a source actually fetched.

Steps

1. Resolve the seed paper (one bash call). $ARGUMENTS is a DOI, arXiv id, OpenAlex id, URL, title, or a path to notes naming a paper.

   SCHOLAR=~/.local/share/anu/plugins/ideas/bin/scholar
   SEED=$("$SCHOLAR" resolve "$ARGUMENTS" --json)   # canonical record + metadata
   STAMP=$(date '+%Y-%m-%d %H:%M')
   TS=$(date +%Y%m%d-%H%M%S)
   

   Sanity-check the resolved title against what the user asked for. If resolution is ambiguous or clearly wrong (a title search can miss, and bare pre-2022 arXiv ids do not resolve), show what resolved and ask before spending the workflow: a dossier on the wrong paper is worse than none. If $ARGUMENTS is an idea with no obvious seed paper, use /prior-art instead, then /study the nearest hit.

2. Run the workflow. Call the Workflow tool with:

   • scriptPath: ~/.local/share/anu/plugins/ideas/skills/study/workflow.js (expand ~ to the real home dir)
   • args: {"seed": <the SEED object, parsed from JSON>, "query": "$ARGUMENTS", "generatedAt": STAMP}

   Wait for it to complete. It returns the dossier as a JSON object.

3. Persist and render (bash + Write). Key the atlas entry on the OpenAlex id (stable), falling back to a title slug:

   ID=$(printf '%s' "$SEED" | python3 -c 'import json,sys; d=json.load(sys.stdin); print((d.get("id") or "paper"))')
   OUT="$HOME/.anu/atlas/papers/$ID"
   mkdir -p "$OUT/history"
   [ -f "$OUT/study.json" ] && mv "$OUT/study.json" "$OUT/history/$TS.json"
   

   Write the returned dossier JSON to $OUT/study.json, then:

   python3 ~/.local/share/anu/plugins/ideas/skills/study/render.py "$OUT/study.json" "$OUT/index.html"
   open "$OUT/index.html"
   

4. Report: one short paragraph with the page path, the closeness verdict (is the idea already done / partly / open and the single nearest paper), and the one or two sharpest openings from the gap. Do not restate the dossier; the page is where it lives.

Rules

• Never edit template.html or render.py per run. The template is fixed by design; all per-paper variation lives in study.json (same contract as /map).
• A search hit is a screen, not a verdict. The workflow screens the nearest candidates cold (fresh context, fetch + quote) before assigning closeness. Do not relabel its verdicts from memory.
• Frame the gap as bounded by what was searched ("nearest is X, differs in Y; found nothing doing Z"), never a blanket "this is novel".
• Old runs accumulate in $OUT/history/. Never delete them; they are the raw material for diffing how a paper's neighborhood and citation graph move over time.
• If scholar returns nothing for a phase (e.g. an arXiv preprint with no parsed references), the page renders that section thin and says so. That is honest, not a failure.