🔬 Research
Elicit Paper Claim Evidence Row Checklist from Library Notes (No Invented Cite Counts)
Compile an Elicit paper-claim-evidence row checklist from pasted library notes only. No invented cite counts, impact ranks, or evidence scoreboards. Not a live Elicit sync.
0Reviews
Prompt
Act as an Elicit paper-claim-evidence row checklist engineer who only uses pasted library notes. You compile a paper-claim-evidence row checklist the notes already support. You do not invent cite counts, impact ranks, evidence scoreboards, or significance claims. This is not a live Elicit sync, not a Consensus invent, and not research consulting advice. You work only from Inputs. Do not invent stats, citations, quotes, URLs, names, IDs, or records that are not in Inputs. Inputs: - Library notes I lock (paper stubs, claim cues, evidence fragments): [LibraryNotes] - Elicit workspace or version notes I lock: [Version] - Review or project label I may quote (or UNKNOWN): [ReviewLabel] - Paper titles already present (or UNKNOWN): [PaperTitles] - Claim cues already present (or UNKNOWN): [ClaimCues] - Evidence cues already present (or UNKNOWN): [EvidenceCues] - Filter cues already present (or UNKNOWN): [FilterCues] - Words I must not use: [Banned] - What I must never invent (cite counts, impact ranks, evidence scoreboards, significance claims): [Never] - Output format: [Format] - Language: [Lang] Generate: 1. Honesty ledger: LibraryNotes nouns, Version, ReviewLabel, PaperTitles, ClaimCues, EvidenceCues, FilterCues, Lang. Banner: not research consulting advice; not a live Elicit sync. Forbidden: invented cite counts, impact ranks, evidence scoreboards, significance claims. 2. Paper-claim-evidence row checklist: one checkbox row per PaperTitles entry. Attach only ClaimCues and EvidenceCues named beside that paper in LibraryNotes. Missing claim or evidence write NOT IN INPUTS. 3. Filter sketch: for each FilterCues entry, list papers that name it. Do not invent a 2400 cite-count claim if absent. 4. Review caution block: quote ReviewLabel and Version only. DOI packs not in LibraryNotes stay NOT IN INPUTS. 5. Refuse list: inventing 2400 cite counts, inventing impact ranks, inventing evidence scoreboards, inventing significance claims. 6. Compliance pass: quote Banned and Never hits. Cut them. Print paper and claim counts from LibraryNotes only. Format as Format. Constraints: - Paper-claim-evidence row checklist from LibraryNotes only. No invented cite counts. - Honor Version. No emojis. Not a live Elicit dashboard. Not research consulting advice.
Instructions
Replace every [bracket] with your details before running. Works on ChatGPT, Claude, and Gemini.
Generated Output
This image was generated using the prompt above.

Examples
Example Input
LibraryNotes: paper title Harbor Quay Diffuser Study as pasted claim cue claim:quay-flow as pasted evidence cue evid:n-of-12 as pasted; paper title Quay Storm Filter Trial as pasted claim cue claim:storm-cut as pasted. Cite invent NONE. Impact invent NONE. Version: Elicit as pasted (do not invent unreleased AI lit coach) ReviewLabel: Cedar Pier evidence bay as pasted PaperTitles: Harbor Quay Diffuser Study as pasted; Quay Storm Filter Trial as pasted. Archive pack UNKNOWN. ClaimCues: claim cue claim:quay-flow as pasted for Harbor Quay Diffuser Study; claim cue claim:storm-cut as pasted for Quay Storm Filter Trial. Abstract pack UNKNOWN. EvidenceCues: evidence cue evid:n-of-12 as pasted for Harbor Quay Diffuser Study; Quay Storm Filter Trial evidence NOT IN INPUTS. FilterCues: filter cue filter:rct-only as pasted for Harbor Quay Diffuser Study. DOI pack UNKNOWN. Banned: 2400 cites, impact rank #1, guaranteed evidence scoreboard Never: invent cite counts, invent impact ranks, invent evidence scoreboards, invent significance claims Format: ledger + paper-claim-evidence row checklist + filter sketch + review caution + refuse + compliance Lang: English
Example Output
1. Ledger. LibraryNotes: paper title Harbor Quay Diffuser Study + claim cue claim:quay-flow + evidence cue evid:n-of-12; paper title Quay Storm Filter Trial + claim cue claim:storm-cut. Cite invent NONE. Impact invent NONE. Version Elicit. ReviewLabel Cedar Pier evidence bay. PaperTitles Harbor Quay Diffuser Study; Quay Storm Filter Trial. ClaimCues claim:quay-flow for Harbor Quay Diffuser Study; claim:storm-cut for Quay Storm Filter Trial. EvidenceCues evid:n-of-12 for Harbor Quay Diffuser Study; Quay Storm Filter Trial evidence NOT IN INPUTS. FilterCues filter:rct-only for Harbor Quay Diffuser Study. DOI pack UNKNOWN. Lang English. Banner: not research consulting advice; not a live Elicit sync. Forbidden: invented cite counts, impact ranks, evidence scoreboards, significance claims, 2400 cites, impact rank #1, guaranteed evidence scoreboard. 2. Paper-claim-evidence row checklist. [ ] Harbor Quay Diffuser Study | claim claim:quay-flow as pasted | evidence evid:n-of-12 as pasted. [ ] Quay Storm Filter Trial | claim claim:storm-cut as pasted | evidence NOT IN INPUTS. Abstract pack not attached. Archive pack not added. 3. Filter sketch. filter cue filter:rct-only | paper Harbor Quay Diffuser Study as pasted. Quay Storm Filter Trial filter | NOT IN INPUTS. Cite counts NOT IN INPUTS so do not invent 2400 cites. Second filter cue not invented. 4. Review caution. ReviewLabel Cedar Pier evidence bay. Version Elicit. DOI pack UNKNOWN. Do not invent significance claim packs. 5. Refuse. 2400 cite counts: refused. impact ranks: refused. evidence scoreboards: refused. significance claims: refused. Unreleased AI lit coach: refused. 6. Compliance. Banned hits none. Papers 2. Claims named 2. Format ledger+paper-claim-evidence row checklist+filter sketch+review caution+refuse+compliance. Gaps: Quay Storm Filter Trial evidence, Quay Storm Filter Trial filter, Abstract pack, Archive pack, DOI pack. 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.