Home / wanshuiyin / anti-autoresearch · GitHub

DESIGN.mdA

DESIGN.md from wanshuiyin/anti-autoresearch is a DESIGN.md file: the design rules coding agents follow in that repo.

Install it, pinned

mdr add wanshuiyin/anti-autoresearch/DESIGN.md@git:20260810.98a75fc

Pins git:20260810.98a75fc in mdr.lock. Pin the fingerprint instead.

Audit grade: A, 14 of 14 checks passed
Kind: DESIGN.md file
Repository: wanshuiyin/anti-autoresearch
Path: DESIGN.md
Versions recorded: 2
Latest version: git:20260810.98a75fc, committed Aug 10, 2026
First version recorded: Jun 25, 2026
Size: 3,394 bytes
Repository stars: 179
License: MIT

Collected from public GitHub. Every version is saved exactly as it was and never changes. Install it at an exact version with the mdr tool, and every file is checked before it reaches your agent. The main file is checked against the SHA-256 fingerprint recorded here, and the others against the fingerprints GitHub keeps for its source commit. The automatic checks below grade the latest version, and the same checks always give the same file the same grade.

What the file says

# Design notes

This repo's architecture was settled by a cross-model design review (the same
discipline the tool enforces): a draft design was put to an independent model
family (codex / gpt-5.5, xhigh) acting as a skeptical architect, and reshaped per
its critique. The four decisions below are the load-bearing ones.

## 1. Positioning: forensics, not detection
The verdict subject is **integrity under limited evidence**, not "is this paper
autoresearch?". Autoresearch is the **failure-prior / taxonomy**, never the thing
being detected. Making "AI-generated" the verdict would collapse the tool into an
AI-text classifier (cluster A) and invite exactly the dismissal we want to avoid.

## 2. The evidence ledger is the spine
Five LLM auditors each parsing a PDF = five hallucinated structures. So a single
**deterministic** pass produces `claims.json` (span-anchored, hashed), and it is
the *only* structure auditors reason over. Invariant: **no span → no high-severity
finding.** This is the difference between "a model said so" and "here is the exact
sentence and its hash".

## 3. The LLM never grades; and neither does the tool
…

Read the whole file at its exact version.

How to install

The exact contents, by fingerprint

mdr add wanshuiyin/anti-autoresearch/DESIGN.md@sha256:7c307d1dbdb19af3

The line at the top of this page pins the label git:20260810.98a75fc, which is what most people want. This one pins the sha256 fingerprint of the contents instead, so it survives a label being moved. Either way the exact version is written to mdr.lock, and mdr install downloads and checks those contents again, so it installs them exactly or stops with an error.

Badge

[![mdr](https://markdownregistry.com/badge/art_xloq2m44d3lyuyzj.svg)](https://markdownregistry.com/a/art_xloq2m44d3lyuyzj)

4 badge image fetches in 30 days. What the badge shows.

Versions

The registry has recorded 2 versions of this file, between Jun 25, 2026 and Aug 10, 2026, about one every 46 days.

version	committed	commit	size	audit
git:20260810.98a75fc latest	2026-08-10	98a75fc	3,394 B	A	view · diff
git:20260625.84dfbfa	2026-06-25	84dfbfa	3,031 B	A	view

Audit of the latest version

A 14 of 14 checks passed. Fixed automatic checks with no AI involved, so the same checks give the same answer every run.

pass: Size between 200 bytes and 200 KB (3394 bytes)
pass: No zero-width or bidi control characters
pass: No instruction hidden inside an HTML comment
pass: No link to an exfiltration or paste host
pass: No credential-shaped string
pass: No instruction to send local credentials anywhere
pass: No text hidden with inline styles
pass: No prompt-injection phrasing
pass: No curl or wget piped into a shell
pass: No recursive delete of root, home or parent
pass: No instruction to read or print local credentials
pass: No base64 blob over 200 characters
pass: No link to a raw IP address
pass: No script tag

Source

GitHub

wanshuiyin/anti-autoresearch · 179 stars · license MIT · pushed 2026-10-30 · branch main

API

GET https://markdownregistry.com/api/v1/artifacts/art_xloq2m44d3lyuyzj
GET https://markdownregistry.com/api/v1/resolve?ref=wanshuiyin/anti-autoresearch/DESIGN.md
GET https://markdownregistry.com/api/v1/blob/7c307d1dbdb19af35cfde1d30385c5de02028b7db247e2bb8d85c8b83e91b114

Your AI assistant does the legwork. You hear about the deals worth your word. Hand yours the standing instructions at modelranch.com and it joins the network that reads files like this one.

Guides: what DESIGN.md is and how agents use it · which agent file does what.

More from wanshuiyin/anti-autoresearch

adversarial-case-builder SKILL.md skill

wanshuiyin/anti-autoresearch · skills/adversarial-case-builder/SKILL.md · Synthesize the single strongest EVIDENCE-BOUND reviewer case to reject a paper, built ONLY from the evidence ledger…