Use when a review looks fabricated and you need to decide between reporting it, replying, or leaving it.
fake-review-assessment.md
Assess whether a review is likely to be fake or policy-violating, and draft the response.
The review: {{REVIEW_TEXT}}
Reviewer profile signals I can see (other reviews, account age, photo, name pattern): {{REVIEWER_SIGNALS}}
Our records for that name, date or job: {{OUR_RECORDS}}
Platform: {{PLATFORM}}
Produce:
1. A signals table with columns: Signal, What I observed, Points towards genuine or fake, Weight (high, medium, low). Only include signals I have actually given you or that are visible in the review text itself.
2. A verdict: likely genuine, unclear, or likely fake or policy-violating. State your confidence and the single strongest signal behind it.
3. If likely fake or policy-violating: which specific platform policy it appears to breach in plain terms (for example no first-hand experience, conflict of interest, off-topic, or personal attack), and a short factual report note under 100 words stating only observable facts.
4. A public reply of 40 to 70 words to post regardless of whether the report succeeds, written for future readers, that neither accuses the reviewer of lying nor accepts the account as true.
Constraints:
- Do not assert that a review is fake as a fact. Use "we have no record of" phrasing.
- Do not guess at the reviewer's identity, employer or motive in public.
- If the signals are genuinely ambiguous, say unclear rather than picking a side to be helpful.
- No em dashes.
Replace each placeholder with your own detail. The more specific you are, the less the model invents.