Rlm Research Analyzer
- Updated April 20, 2026
- sfoucher/rlm-research-analyzer
RLM Research Analyzer is a Claude Code skill that synthesizes Zotero collections into verified literature reviews using recursive worker delegation so developers can analyze large paper sets without context saturation.
About
rlm-research-analyzer is a Claude Code marketplace plugin that turns a Zotero collection into a structured literature review using a Recursive Language Model workflow. A root agent indexes metadata, plans research questions, spawns parallel Haiku workers to summarize paper batches, aggregates themes and contradictions, and runs adversarial verification before delivering Markdown, PDF, and Zotero notes. Use it when a paper collection is too large for one context window and you need cited synthesis with confidence tags rather than a single-pass summary.
- Six-stage pipeline: initialize, plan, delegate, aggregate, verify, deliver
- Parallel Haiku workers read Zotero full text via zotero-mcp per batch
- Adversarial reviewer runs twice fixing FATAL and MAJOR issues before delivery
- Outputs Markdown synthesis, PDF, provenance file, and per-paper Zotero notes
- Optional semantic filtering when a focus question and embedding API are configured
Rlm Research Analyzer by the numbers
- Data as of Jul 7, 2026 (Skillselion catalog sync)
/plugin marketplace add sfoucher/rlm-research-analyzerAdd your badge
Show developers this marketplace is listed on Skillselion. Paste this into your README.
| Last updated | April 20, 2026 |
|---|---|
| Repository | sfoucher/rlm-research-analyzer ↗ |
How do I produce a rigorous literature review from a large Zotero collection when no single model pass can read every paper's full text?
Synthesize a Zotero paper collection into a verified literature review using recursive delegation to parallel Haiku workers and adversarial review.
Who is it for?
Researchers and developers with Zotero plus zotero-mcp who need structured, adversarially reviewed literature synthesis in Claude Code.
Skip if: Users without Zotero desktop and zotero-mcp configured, or anyone needing quick summaries without planned research questions and verification passes.
What you get
Writes rlm_index, rlm_plan, slice files, rlm_answer Markdown and PDF, rlm_provenance metadata, and creates a Zotero note on each processed paper.
Plugins in this marketplace
1 plugin - install individually after you add the marketplace.
Recommended Marketplaces
FAQ
What prerequisites are required?
You need Claude Code, Zotero desktop running locally, and the zotero-mcp server configured in settings.json.
How does the RLM pattern map to this plugin?
The root sees metadata only, Haiku workers read batched full text into slice files, the root aggregates findings, and an adversarial reviewer validates the synthesis.
What deliverables does a run produce?
Final outputs include rlm_answer Markdown, an optional PDF, provenance metadata, intermediate slice and review files, and per-paper Zotero notes.