Citation Sanity Check
A step-by-step method for verifying references and quotes produced by a model.
The explanation is in the selected language; the prompt text stays in English.
Why use it
For anyone who must vet model-written text or references before using them: a skeptical librarian tells you what to check and where for each item.
How to use it
Paste text with references or claims into [PASTE]. The model verifies nothing; it only gives a checklist you then follow yourself on the publisher page or in a database.
Short example
Text: Smith et al. (2019) in The Lancet showed this method reduced mortality; no DOI given.
What to expect
A table with item, what to verify, where and an initial suspicion level. Every row is labeled UNVERIFIED until I check it myself.
Precautions and tips
- The model cannot browse; do not trust even a reference it marks plausible.
- Match authors, year and journal through PubMed or Crossref.
- Keep unchecked citations out of your work.
# Citation Sanity Check
## Role
You are a skeptical librarian.
## Context
Text with references or claims to verify: [PASTE]
## Task
For each reference or factual claim, (1) state what exactly should be checked, (2) suggest where to check it (publisher page, database, original document), (3) mark whether it looks plausible or suspicious based only on internal consistency (names, years, journal fit).
## Output format
A checklist table: item, what to verify, where, initial suspicion level.
## Constraints
- You cannot browse; never claim a reference is verified.
- Label every item 'UNVERIFIED until I check it myself'.The prompt text is the original English and is not translated: paste it into your AI tool as is. Replace [text in square brackets] or CAPITALIZED placeholders with your own details. Always check the answer.
Interactive mode: search, progress and Python exercises.