菜单
PubMed + bioRxiv search for recent literature on the patient's genes and diagnosis. bioRxiv preprints are first-class (earliest signal). Emits Tier C; the LLM only summarizes genuine free text, it never asserts facts.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Open Targets gene-disease association lookup via the official MCP connector. One GraphQL round trip per gene. Emits Tier C informational evidence.
Standing watcher. Reruns the trial, literature (PubMed + bioRxiv), and ClinVar-relookup specialists against the FIXED profile and diffs against the stored baseline. Empty diff -> one calm "no significant changes" timeline entry.
ClinicalTrials.gov v2 search for trials relevant to the patient's diagnosis and genes. Eligibility is surfaced as "may be eligible — confirm with your care team", never confirmed. Emits Tier C.
Uncertainty-ranked gap report (spec §8b). For an unresolved (Tier B) variant, names that the analysis can't resolve it and attributes the gap to a missing input — a documented family history — that would let the analysis weight it. Never a test order.
Uncertainty-ranked gap report (spec §8b). Names a pharmacogenomic module the analysis could NOT complete and attributes the gap to a specific missing input (a genotype). Never tells the patient to get a test — it is a report on the tool's blind spots for the care team.
Deterministic CPIC pharmacogenomics lookup over the patient's current medications. Pure code, no LLM. Emits Tier A/B for CPIC Level A/B gene-drug pairs; flags a missing genotype instead of ever directing a dose change.
| name | literature-synthesis |
| kind | oneshot |
| domain | literature |
| description | PubMed + bioRxiv search for recent literature on the patient's genes and diagnosis. bioRxiv preprints are first-class (earliest signal). Emits Tier C; the LLM only summarizes genuine free text, it never asserts facts. |
Search PubMed (E-utils, use mindate) and bioRxiv (use since) for recent
literature on the profile's genes and diagnosis. bioRxiv is first-class here — the
developer specifically wants preprint-level early signal (spec §8a).
Emit Tier C EvidenceRecords. The LLM's only job is to summarize genuinely
free-text abstracts into a minimally-reworded claim; it does not assert medical
facts, and every rendered string passes the code gate. source_id is the PMID or
bioRxiv DOI; source_url resolves to the article. Never invent a PMID (spec §3) —
if a source can't be resolved, drop the claim.
STATUS: live via app/literature.py — peer-reviewed through NCBI E-utils and
preprints through Europe PMC (SRC:PPR, which covers bioRxiv). No MCP/auth needed.
Runs only when BWCLS_LIVE_SOURCES=1. Emits Tier C records; titles become the
(factual) claim and are gate-scanned. since= narrows to new-only for the recheck;
world-evidence-watch reuses search_literature.
TODO (optional): LLM abstract summarization (still gate-checked) instead of titles; or switch to the developer's PubMed/bioRxiv MCP connectors.
