/agentsop-bio-fraud-forensics

name: agentsop-bio-fraud-forensics
domain: research-integrity
trigger_keywords:
  - "data fraud / image manipulation"
  - "Western blot duplication / splicing"
  - "GRIM / statcheck / impossible statistics"
  - "paper mill / tortured phrases"
  - "PubPeer / Retraction Watch verification"
description: >-
  Screens biomedical / life-science papers for signs of data fabrication, image
  manipulation, and statistical anomalies, using the detection techniques distilled
  from the field's canonical exposure platforms (PubPeer, Data Colada, Science
  Integrity Digest, For Better Science) and tools (ImageTwin/Proofig, statcheck,
  GRIM/GRIMMER, Problematic Paper Screener, Seek & Blastn). Use when asked to check
  a paper/figure for image duplication, blot splicing, impossible statistics, paper-mill
  or tortured-phrase signals, research integrity, or "is this data faked"; or when a
  user shares a figure, Western blot, supplementary dataset, or DOI and asks whether it
  looks manipulated. Reports observable anomalies as questions for clarification — it
  never accuses anyone of fraud.
version: 1.0.0

Bio-Fraud Forensics · 生物医学论文数据造假筛查

A screening methodology for life-science papers. It reverse-engineers how real cases were caught — the exact panels compared, the transform applied, the statistic recomputed — and turns that into a reproducible per-paper checklist. It is a **detective's lens, not a verdict machine**: every output stays at "observed anomaly" or "question for the authors," because red flag ≠ proof and an accusation can end a career.

Activation Rules

**Trigger when:**

• "Check this paper / figure / Western blot for manipulation," "does this data look faked," "screen for image duplication."
• A user shares a figure, blot, microscopy panel, supplementary `.xlsx`, or a DOI and asks if it's trustworthy.
• "Is this a paper mill?", "tortured phrases," "are these statistics possible," "run GRIM/statcheck on this."
• "Where do I check if this paper has been flagged / retracted?" (verification routing).
• Asked to draft a PubPeer-grade, reproducible image/data integrity comment.

**Do NOT trigger when:**

• The user wants a scientific peer review of validity/novelty (use a peer-review skill) rather than an integrity screen.
• The user asks you to publicly accuse a named person of fraud, or to write an accusation/social post (refuse — see Boundary Rules).
• The task is general statistics help or figure-making with no integrity question.
• The paper is non-biomedical and the request is about a domain whose fraud signatures differ (physics/CS); say so and scope down.

Agentic Protocol

Run this as a chain-of-steps. Cheapest, fastest signals first; the expensive image/stat forensics last (they tell you *where* to dig is often answered for free by the cheap checks).

**Step 1 — Scope & status.** Identify the input: single figure, full paper, supplementary dataset, or a batch. Run the status cascade in parallel (it's free and may hand you the answer): Retraction Watch Database → PubMed retraction banner → Crossref/Crossmark notice → PubPeer (search DOI/author) → ORI case index (only if adjudicated US PHS misconduct is the question). Note what already exists; your job may shift to verifying/extending a prior flag.

**Step 2 — Ordered screen.** Walk the pipeline, recording each hit; do not stop at the first: 1. *Metadata/affiliations* — email domains, ORCID freshness, affiliation vs claim, special-issue venue. 2. *Text-mechanical* — tortured phrases ("bosom peril"=breast cancer), LLM leakage ("as an AI language model"), recycled/irrelevant references. 3. *Image forensics* (the #1 biomedical signal) — see M2; classify each duplication Bik Type I/II/III. 4. *Statistical forensics* — see M3; GRIM/GRIMMER/statcheck/SPRITE + digit/uniformity; `.xlsx` → calcChain. 5. *Raw-data availability* — are uncropped originals / source data provided and openable? 6. *References integrity* — do sampled citations resolve and support the claim? For stats-heavy/clinical papers, swap 3 and 4. For a *batch* question, run M5 (recurrence across papers is the signal).

**Step 3 — Match a model & classify.** For each hit, Read `references/sop_models.md`, match the operation model (M1–M7), and name the sub-type + Bik category. Confirm image matches by performing the transform yourself (flip/rotate/overlay) and including the result; confirm any tool flag by human inspection — a large share of automated image hits are benign reuse, so treat none as a finding until you have reproduced it by hand.

**Step 4 — Benign-explanation gate (mandatory before any escalation).** Run the benign-explanation checklist in M6. Record which innocent causes were excluded and why (disclosed splice, JPEG block, same-experiment loading-control reuse, tiling overlap, figure-assembly slip). No "looks suspicious → flag." Apply the honest-error discriminators from M1 (directionality, recurrence, sophistication, provenance, disclosure).

**Step 5 — Grade & document.** Default every finding to **Tier 1 (observed anomaly)**. Escalate to **Tier 2 (question for authors)** only after Step 4, using the disclosed-evidence + hedge + named-alternative formula. Never originate **Tier 3 (adjudicated misconduct)** — cite the body that ruled. Write each finding in the reproducible annotation format (M7) and pick an Output Mode.

Core Operation Models

| # | Model | Core proposition | Main source | |---|-------|------------------|-------------| | M1 | **FFP Taxonomy & Honest-Error Discriminators** | Classify the anomaly (fabrication/falsification + sub-types); separate honest error from misconduct via 5 tests; only ever assert the "significant departure," never intent. | ORI/42 CFR 93; Bik mBio 2016 | | M2 | **Image Forensics** | Every band/field is a fingerprint; catch by eye, confirm by flip/rotate/overlay-Difference; correlated *background* texture (not band shape) is decisive; Bik Type I/II/III drives escalation. | Bik; ASM/ImageTwin pilot; Proof