The new Amass app is live. Ask in plain language, check every claim. Try it free

Skills & PromptsSubmission evidence assembler

Skill

Submission evidence assembler

Name one drug asset; get a citation-auditable, Module-2.5-style evidence narrative — every pivotal Phase-3 trial, its describing papers, and a canonical Amass-ID audit trail — assembled from Amass's publication↔trial graph with trust metadata on every cited paper.

New to Amass? Connect Amass to Claude first

6Phase-3 trials
37linked papers
0retracted
You getxlsxdocx2 files
Skillregulatoryclaude

The prompt

Paste this into Claude

Asset (drug)mavacamten
prompt
Use the submission-evidence-assembler skill.

Asset (drug): mavacamten

Why not just ask the model?

Ask a plain model to "assemble the published evidence base for mavacamten's HCM submission" and it invents the asset→trial→paper linkage — plausible trial acronyms, PMIDs, journals, and citation counts that read like the literature and are wrong.

  • Plain model: fabricated trial acronyms, made-up NCTs and PMIDs, guessed citation counts, and no way to tell a retracted paper from a sound one.
  • Amass: the real TrialCore rows for the asset's Phase-3 search, and each trial's referencesBiomedCore edge walked in one hop to the publications that describe it — every paper arriving with its PMID, journal, journalQualityJufo tier, isRetracted flag, and citation count copied verbatim. The cross-core edge is the load-bearing primitive; the canonical Amass IDs (AMTC_… / AMBC_…) are the stable audit trail.

Why a skill, not a prompt?

This is a multi-step orchestration, not a single question: one TrialCore search for the asset's pivotal trials, a completeness check of that result set against the search limit, a get_amass_trialcore_record per trial to read its identity and referencesBiomedCore array, then a fan-out — a get_amass_biomedcore_record for every referenced paper (37 fetches here) — followed by a union/dedupe across trials and a two-file Office assembly. Packaging it as a skill means you trigger the whole chain with one line instead of pasting a brittle, branch-heavy procedure every time.

Honest scope

It assembles and trust-screens evidence; it does not author regulatory conclusions. The Phase-3 search returns at most limit trials, 10 by default — for mavacamten it returned 6, so the trial set is complete and the skill says so; an asset that fills the limit is a relevance sample, not a census, and the skill fires a truncation banner (broaden by widening phase or wording and unioning the results). Per-paper figures are verbatim Amass fields and can lag the registry by hours; any abstract-derived claim the returned text does not state is written "not in abstract," never guessed. The verdict "0 of 37 cited papers carry isRetracted=true" is a checkable field claim, not an assurance of quality.

Install the skill

Add the skill to Claude

Install every Amass skill with one command, npx skills add amass-technologies/public-skills or download the SKILL.md from GitHub and add it to Claude (Settings → Capabilities → Skills, or drop it in your Claude Code skills directory). Then paste the prompt above to trigger it.

More from the library

Keep exploring

xlsx
14trials watched
Short nameSource idRegistryCopy of
Acute, 50/75 mg vs placeboNCT04072354ClinicalTrials.gov
↳ EUCTR2019-000470-36-BGEU CTRNCT04072354
Acute, 75/100 mg vs placeboNCT04092686ClinicalTrials.gov
competitive-intelligenceSkill

Watchlist Monitor

Name the trials, papers, approvals or drugs you need to follow; get one Excel board that re-reads them through Amass at every check and tells you exactly what changed: status, dates, enrollment, results posted, label revisions, retractions, new links.

pdf+1
196trials in scope
Landscape Monitor — page 1
competitive-intelligenceSkill

Landscape Monitor

Name a field; get it mapped across trials, papers, patents, drugs, genes and approvals into a landscape page with charts, a workbook and a one-page briefing, every record screened with its reason. Ask for an update later to see what is new and what changed.

csv
12targets scored
kinasekinase_family_chembl_leafsmall_molecule_tractabilityprotac_tractability
PLK1Other protein kinase PLK familyAdvanced Clinicalpredictive-only
WEE1Other protein kinase WEE familyAdvanced Clinicalpredictive-only
CDK1CMGC protein kinase CDK5 subfamilyAdvanced Clinicalpredictive-only
researchSkill

Target prioritization matrix

A general, interview-driven target-discovery workbench: name the biology and the modality you care about, and get back a scored table of candidate targets — druggability, essentiality, genetic constraint, curated safety, and the real number of drugs already targeting each — assembled from GeneCore intelligence and DrugCore/TrialCore/BiomedCore/PatentCore cross-links.

Run it yourself

Connect Amass to Claude and paste the prompt.

The connector takes about a minute to add, then this workflow runs on your own inputs.

Data
No training on your data
Hosting
EU-hosted, enterprise
Isolation
Tenant-isolated by architecture
Certification
ISO 27001, SOC 2, NIS2: in motion