skills.rest
mcleanT
mcleanTCOMMUNITY
Explore all skills by this author →
GitHub

run-pipeline

Run the AutoReview pipeline locally via Claude Code CLI to generate scientific review papers.

Updated Feb 24, 2026
Search Tags:#pydantic#cli#pipeline#snapshot#literature-review#autoreview#claude_code
One-click install
npx skills add https://github.com/mcleanT/AutoReview --skill run-pipeline-mcleant
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: run-pipeline
Source: https://github.com/mcleanT/AutoReview/tree/main/.claude/skills/run-pipeline
Command: npx skills add https://github.com/mcleanT/AutoReview --skill run-pipeline-mcleant

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Provides a reliable, production-grade way to run the entire AutoReview literature-to-review pipeline locally with LLM calls routed through the Claude Code CLI, preserving full-texts, in-memory state, validation, and critique loops so outputs match server-side fidelity without manual orchestration.

Core Features & Use Cases

  • Run the real Python pipeline end-to-end (DAGRunner, PipelineNodes, KnowledgeBase) with LLM calls proxied to the Claude Code CLI.
  • Preserve full texts and in-memory KnowledgeBase, enable Pydantic validation of LLM responses, deterministic citation resolution, and automated critique loops for iterative quality control.
  • Support long-running executions with snapshot-based resume, background runs, per-node progress reporting, and configurable settings (domain, depth, date range, model).
  • Use cases include generating publication-ready review papers, resuming interrupted runs from snapshots, evaluating outputs with the claude_code provider, and converting final drafts to preferred formats.

Quick Start

Run the AutoReview pipeline for the topic "the role of gut microbiome in neurodegenerative diseases" using the claude_code provider and confirm the presented configuration to launch.

Frequently Asked Questions about run-pipeline

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I generate a scientific literature review paper locally using Claude?▼

You can generate a scientific literature review paper locally by running the AutoReview pipeline, which routes LLM calls through the Claude Code CLI to handle literature search, extraction, and synthesis while maintaining full Python pipeline fidelity.

Can I resume an interrupted literature review generation pipeline from a previous state?▼

Yes, you can resume interrupted literature review generation using snapshot-based resume functionality, which restores the in-memory KnowledgeBase and pipeline state so long-running background executions can continue without data loss.

How does Pydantic validation work for LLM responses in a research pipeline?▼

Pydantic validation enforces structured schema compliance on LLM responses during the pipeline, ensuring that extracted literature data, citations, and synthesized text meet deterministic quality standards before progressing to the next pipeline node.

What is the best way to run an automated critique loop for a research paper draft?▼

The best way to run automated critique loops is by executing the AutoReview pipeline locally, which iteratively evaluates and refines publication-ready drafts using LLM calls proxied through the claude_code provider for quality control.

Does the AutoReview pipeline support configurable settings for research domain and depth?▼

Yes, the AutoReview pipeline supports configurable settings including research domain, depth, date range, and model selection, allowing you to tailor the literature search, extraction, and synthesis parameters to your specific topic requirements.

Why use the Claude Code CLI instead of direct API calls for a literature review pipeline?▼

Using the Claude Code CLI preserves full-texts, in-memory state, and deterministic citation resolution while matching server-side fidelity, eliminating the need for manual orchestration of complex literature search and critique loops.

Related Skills

Explore More Education & Research Skills→
tblakex01tblakex01
1

setup-gbrain

Install the gbrain CLI, initialize a local brain, and register MCP.

Community
Advanced
exiaoexiao
37

substack-draft

Save finalized article drafts to a Substack publication for manual review.

Community
Intermediate
VibeTechnologiesVibeTechnologies
1

knowledgebase-search

Retrieves indexed knowledgebase documents with BM25 and keyword fallback for answering questions.

Official
Intermediate
blacklanternsecurityblacklanternsecurity
253

ssti-twig

Exploit Twig/PHP server-side template injection to achieve remote code execution.

Official
Advanced