# paperzilla

> Chat with your agent about projects, recommendations, and canonical papers in Paperzilla — covering recent project recommendations, canonical paper details, markdown-based summaries, recommendation feedback, and feed export.

**Use case**: Chat with Paperzilla for curated paper recommendations

**Canonical URL**: https://agentcookbooks.com/skills/paperzilla/

**Topics**: claude-code, skills, science, scientific-writing

**Trigger phrases**: "recommend papers from Paperzilla", "get my Paperzilla feed", "find canonical papers on", "paperzilla recommendations"

**Source**: [K-Dense AI](https://github.com/K-Dense-AI/scientific-agent-skills/tree/main/scientific-skills/paperzilla)

**License**: MIT

---

## What it does

`paperzilla` is a Claude Code skill from K-Dense AI's [scientific-agent-skills repo](https://github.com/K-Dense-AI/scientific-agent-skills). It turns Claude into a Paperzilla interface agent that surfaces curated paper recommendations, canonical papers for a topic, and your personalized feed from [Paperzilla](https://paperzilla.ai) — formatted as markdown summaries or Atom feed exports.

A session produces a curated reading list: recommended papers with summaries, links, and relevance context, or a structured feed export you can import into a reference manager.

## When to use it

Reach for it when:

- You want curated, high-quality paper recommendations for a research area rather than raw database search results
- You're looking for the canonical / foundational papers on a topic that Paperzilla's editorial layer has surfaced
- You want to export your Paperzilla feed for offline reading or import into Zotero

When *not* to reach for it:

- Exhaustive systematic searches across all literature — use `paper-lookup` or `literature-review`
- Direct API access to PubMed, arXiv, or Semantic Scholar — use `paper-lookup`

## Install

Copy the `SKILL.md` from K-Dense AI's [paperzilla folder](https://github.com/K-Dense-AI/scientific-agent-skills/tree/main/scientific-skills/paperzilla) into `.claude/skills/paperzilla/` in your project. A Paperzilla account is required for personalized recommendations.

Trigger phrases: "recommend papers from Paperzilla", "get my Paperzilla feed", "find canonical papers on".

## What a session looks like

A typical session has three phases:

1. **Topic or feed query.** Specify a research topic or ask for your recent Paperzilla feed. Claude formats the request to the Paperzilla interface.
2. **Recommendation retrieval.** Claude surfaces recommended papers with markdown summaries — title, authors, venue, abstract excerpt, and Paperzilla's relevance signal.
3. **Feedback or export.** You can provide feedback on recommendations to tune future suggestions, or export the current set as an Atom feed or formatted list for your reference manager.

## Receipts

**Where it works well:**
- Finding canonical papers in established fields where Paperzilla's curation has had time to identify high-impact works
- Getting a curated reading list for a new research area without having to construct database queries from scratch

**Where it backfires:**
- Very recent or niche topics with limited Paperzilla coverage — the curation layer needs time and community signal to surface quality papers
- Users without a Paperzilla account get generic rather than personalized recommendations

**Pattern that works:** use Paperzilla for the reading queue and `paper-lookup` for targeted retrieval; the two tools complement rather than replace each other.

## Source and attribution

Skill authored by [Paperzilla Inc.](https://paperzilla.ai). The canonical SKILL.md lives in the [`paperzilla` folder](https://github.com/K-Dense-AI/scientific-agent-skills/tree/main/scientific-skills/paperzilla) of K-Dense AI's scientific-agent-skills repository.

License: MIT. Install, adapt, and redistribute with attribution preserved.

This page documents the skill from a practitioner's perspective. For the formal spec and any updates, defer to the source repo.