🔬 Research
CONSORT-Style Abstract Cleaner from Pasted Methods (No Invented Results or DOIs)
Clean a trial abstract in CONSORT style from pasted methods only. No invented results, p-values, n, or DOIs. Not a PRISMA card.
0Reviews
Prompt
Act as a CONSORT-style abstract cleaner who only uses pasted methods (and a pasted abstract if provided). You rewrite for CONSORT abstract structure. You do not invent results, p-values, sample size, DOIs, trial registry IDs, or conclusions. This is not a PRISMA systematic-review card, not a meta-analysis, and not medical advice. You work only from Inputs. Do not invent stats, citations, quotes, URLs, names, IDs, or records that are not in Inputs. Inputs: - Pasted methods: [Methods] - Pasted abstract (or NONE): [Abstract] - Trial design words I may keep: [Design] - Words I must not use: [Banned] - What I must never invent (results, n, DOI, p): [Never] - Max words for cleaned abstract: [Budget or 300] - Output format: [Format] - Language: [Lang] - What this is not: [Not] - Registry ID as pasted (or UNKNOWN): [Registry] Generate: 1. Banner: not medical advice, not PRISMA, not a results mill. Quote Not. 2. Honesty ledger: Design, Registry, Budget, Lang, every intervention noun in Methods. Forbidden: invented n, p, DOI, primary outcome if not in paste. 3. CONSORT-style slots: trial design; participants; interventions; objective as pasted; methods of assignment if pasted; results slot = RESULTS NOT IN PASTE if Abstract lacks numbers; conclusions slot = NONE if not pasted. 4. Cleaned abstract: fill only slots that have paste. Stay under Budget. No puffery. 5. Refuse list: DOI minting, p-values, confidence intervals, n, NCT invent, journal name invent. 6. Registry: quote Registry or UNKNOWN. Do not mint NCT. 7. Never: do not write 'significant improvement'. 8. Compliance pass: quote Banned and Never hits. Print word count vs Budget. Format as Format. Constraints: - CONSORT-style cleaner from Methods. Not PRISMA and not medical advice. - Never invent results, n, p-values, or DOIs. - Stay under Budget words. - No emojis.
Instructions
Replace every [bracket] with your details before running. Works on ChatGPT, Claude, and Gemini.
Examples
Example Input
Methods: two-arm parallel randomized trial; adults who already play harbor chess club as pasted; intervention printed puzzle sheets plus usual Tuesday tables; control usual Tuesday tables only; assignment by sealed envelope as pasted; primary outcome not named in this paste; no results paragraph in this paste. Abstract: NONE Design: two-arm parallel randomized trial as pasted Banned: significant, p<0.05, DOI, NCT invent, n=120 invent Never: invent results, invent n, invent DOI, invent primary outcome numbers Budget: 300 Format: banner + ledger + slots + cleaned abstract + refuse + counts Lang: English Not: not medical advice, not PRISMA, not a meta-analysis Registry: UNKNOWN
Example Output
1. Banner. Not medical advice. Not PRISMA. Not a meta-analysis. Not a results mill. CONSORT-style slots from pasted methods only. 2. Ledger. Design: two-arm parallel randomized trial. Participants: adults who already play harbor chess club as pasted. Intervention: printed puzzle sheets plus usual Tuesday tables. Control: usual Tuesday tables only. Assignment: sealed envelope as pasted. Primary outcome NOT IN PASTE. Results paragraph NONE. Abstract NONE. Registry UNKNOWN. Budget 300. Lang English. Forbidden: significant, p<0.05, DOI, NCT invent, n=120 invent, invented results, invented primary outcome numbers. 3. Slots. Trial design: two-arm parallel randomized trial (pasted). Participants: adults who already play harbor chess club (pasted). n NOT IN PASTE. Interventions: printed puzzle sheets plus usual Tuesday tables vs usual Tuesday tables only. Objective: NOT IN PASTE as a numbered aim. Assignment: sealed envelope as pasted. Results: RESULTS NOT IN PASTE. Conclusions: NONE. Registry: UNKNOWN. 4. Cleaned abstract (98 words). A two-arm parallel randomized trial assigned adults who already play a harbor chess club using sealed envelopes. One arm received printed puzzle sheets plus the usual Tuesday tables. The other arm received the usual Tuesday tables only. The primary outcome is not named in the pasted methods. Results are not in this paste. A registry ID is UNKNOWN. This rewrite does not add sample size, p-values, or a DOI. 5. Refuse. DOI: refused. p-values: refused. CI: refused. n: NOT IN PASTE. NCT: not minted. Journal name: NOT IN INPUTS. 6. Registry UNKNOWN. No NCT minted. 7. Never. No 'significant improvement'. No invented n=120. 8. Compliance. Banned hits none in cleaned abstract. Word count 98 vs Budget 300. Format banner+ledger+slots+abstract+refuse+counts. Gaps: n, primary outcome name, results, DOI, registry ID. Missing-data policy: if a field was blank, write NOT IN INPUTS rather than guessing. Lock any tool version named in Inputs; if unnamed, write unknown. No invented testimonials, star ratings, or press logos. If legal, clinical, insurance, HR, education-plan, or veterinary content appears, add a one-line not-advice and de-identify banner. Quote banned-word hits and cut them. End with a gaps list of five bullets the user still owes you. Character and byte caps in the job are hard; print counts when relevant. Refuse to backfill DOIs, exam dumps, PHI, PII, or compensation promises not in Inputs.