corpus

Analyze large document corpora with multi-pass synthesis and citation provenance.

Updated Jun 13, 2026
One-click install
npx skills add https://github.com/Seth090502/osanwe-public --skill corpus
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: corpus
Source: https://github.com/Seth090502/osanwe-public/tree/main/.claude/skills/corpus
Command: npx skills add https://github.com/Seth090502/osanwe-public --skill corpus

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires Claude Code, subagents, and includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve?

This Skill facilitates systematic analysis of multi-document corpora (10-150 documents), providing verifiable multi-pass synthesis with citation provenance.

Core Features & Use Cases

  • End-to-end Analysis: Analyze and synthesize large document corpora with verifiable multi-pass synthesis.
  • Citation Discipline: Ensures every claim is cited and evidential class tagging is correct.
  • Adversarial Review: Includes adversarial review at multiple stages to maintain accuracy.
  • Use Case: Use this Skill for FOIA archives, declassified records, legal discovery sets, regulatory filings, court records, etc., to systematically analyze documents with provenance and verifiability.

Quick Start

Run the following command to initiate the analysis: /corpus init --name <slug> --source <source-dir>

Frequently Asked Questions about corpus

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

FAQPage Schema
How do I systematically analyze large document corpora for legal discovery?

Systematically analyze large document corpora by applying multi-pass synthesis and citation provenance. This approach processes 10-150 documents, ensuring every claim is cited and evidential class tagging is correct for legal discovery sets.

What is multi-pass synthesis for FOIA archives and declassified records?

Multi-pass synthesis for FOIA archives is an analysis mechanism that iteratively reviews declassified records to extract verifiable claims. It includes adversarial review at multiple stages to maintain accuracy and structured output for provenance tracking.

Can I process regulatory dockets and court filings with Claude Code subagents?

Yes, you can process regulatory dockets and court filings using Claude Code with subagents. The Skill enforces model floor requirements and leverages subagents to handle structured output and provenance tracking for mass-document analysis.

How do I start corpus analysis on a directory of intelligence community releases?

Initiate corpus analysis on intelligence community releases by running the command `/corpus init --name <slug> --source <source-dir>`. This sets up the environment to systematically analyze the documents with verifiable multi-pass synthesis.

What is the maximum document limit for provenance tracking in corpus analysis?

The maximum document limit for provenance tracking in corpus analysis is 150 documents. This Skill is designed to systematically analyze and synthesize multi-document corpora ranging from 10 to 150 documents.

Does document analysis ensure citation discipline for every extracted claim?

Yes, document analysis ensures citation discipline for every extracted claim. The Skill enforces strict citation provenance and evidential class tagging, incorporating adversarial review at multiple stages to maintain accuracy.