Back to portfolio

case-study

Research Claim Ledger

ClawHub skill-build case study for turning research drafts, source packets, literature matrices, or reviewer notes into source-backed claim ledgers.

Overview

Research Claim Ledger is a narrow public ClawHub skill for academic and technical writing review. It avoids the over-broad promise of a full research suite and focuses on one trust-building artifact: a claim-by-claim ledger that marks each sentence as supported, weakly-supported, overclaimed, wrong-source, missing-locator, stale-source, inaccessible, unsupported, or needing human review. The public case study compares this narrower wedge against broad academic-agent workflows and frames the deliverable as a receipt a writer can share with a supervisor, coauthor, or reviewer.

What It Covers

  • Turns draft sections, literature matrices, notes, citation lists, source packets, or reviewer comments into a structured claim ledger
  • Separates supported claims from overclaims, missing locators, stale sources, inaccessible references, and human-review cases
  • Keeps the first release instruction-only with no bundled scraping helper or hidden dependency surface
  • Positions the skill as a small source receipt rather than a paper-writing, plagiarism, legal, medical, or financial review tool

Stack And Topics

  • ClawHub
  • Codex Skills
  • Research QA
  • Citation Review
  • Markdown
  • Release Engineering

Public Signals

  • ClawHub downloads: 198 public ClawHub listing, 2026-06-04
  • Published versions: 1 public ClawHub listing, 2026-06-04
  • Verdict labels: 9 supported, weakly-supported, overclaimed, wrong-source, missing-locator, stale-source, inaccessible, unsupported, needs-human-review
  • Release posture: instruction-only SKILL.md and agent config first release; scripts deferred until example ledgers validate the workflow

References