Skip to content
Get packs

Integrity Forensics Paper Audit

Runs a deterministic statistical-integrity check on a paper and issues a block/warn/pass gate.

Source Download skill
Repository
wanshuiyin/Auto-claude-code-research-in-sleep
Stars
18,482
Licence
MIT
Files
1 file

What it does

When installed, the agent can audit a research paper (folder, PDF or arXiv ID) for numerical integrity problems such as inconsistent statistics (GRIM, GRIMMER, statcheck) and arithmetic deltas. It clones a pinned version of an external tool, runs its rule-based checks, and turns the result into a gate labelled BLOCK, WARN or NO_NEW_BLOCKER, plus a ledger that records findings and how they were resolved. It cannot give a paper a clean bill of health; in this Codex version the semantic review dimensions do not run.

Great for

  • Authors doing a self-check on statistics before journal submission
  • Reviewers or students who want a quick first pass for numeric inconsistencies
  • Teams wanting a repeatable, logged integrity checklist for papers

Skip it if

  • Final integrity certification: the skill can flag but never returns CLEAN, and its semantic checks are not run in this mode, so it misses many issues.
  • Auditing papers you do not have permission to process, since it clones and runs code and may send data over the network.
  • Situations needing an automatic pass: the gate is designed to block when no flags are found, so it will not wave papers through.

Try asking

  • “Run integrity forensics on ./my-paper and tell me what it flags”
  • “Forensic audit this paper before I submit: /path/to/manuscript”
  • “投稿前自查诚信 for the PDF in ./draft.pdf”
  • “Check the statistics in arXiv 2401.00001 for GRIM inconsistencies”
Step-by-step guide: how to use Integrity Forensics Paper AuditInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 1 original files, the MIT licence and an attribution note.

    integrity-forensics.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip integrity-forensics.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip integrity-forensics.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • Point it at the paper directory with absolute paths so the upstream tool can locate files.
  • Treat any HARD_FLAGS or REVIEW_UNAVAILABLE result as a to-do list; fix the numbers in the source data rather than re-running until the flags disappear.
  • For the full nine-dimension semantic review, use the Claude Code version of the skill instead.

What's inside

  • SKILL.md

Redistributed unmodified from wanshuiyin/Auto-claude-code-research-in-sleep under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.