Loading
Retrieves ClinGen gene-disease validity assertions for a public gene or disease, and reviews source-linked public evidence and literature for one supported GRCh38 germline nuclear SNV or simple indel
The skill file stays in the publisher's repository.
Paste the copied command in a terminal. It installs this SKILL.md into that agent.
npx skills add K-Dense-AI/scientific-agent-skills --skill folklore-variant-evidence -a cursor -y
Retrieves ClinGen gene-disease validity assertions for a public gene or disease, and reviews source-linked public evidence and literature for one supported GRCh38 germline nuclear SNV or simple indel through Folklore Clinical Variant Interpretation MCP. Used when a scientific agent must branch deterministically on resolved, ambiguous, not-found, invalid, unsupported, or unavailable variant outcomes; chain a resolved public variant into related literature or publication details; or preserve evidence provenance without accepting patient, phenotype, family, segregation, or private case data.
Retrieves ClinGen gene-disease validity assertions for a public gene or disease, and reviews source-linked public evidence and literature for one supported GRCh38 germline nuclear SNV or simple indel through Folklore Clinical Variant Interpretation MCP. Used when a scientific agent must branch deterministically on resolved, ambiguous, not-found, invalid, unsupported, or unavailable variant outcomes; chain a resolved public variant into related literature or publication details; or preserve evidence provenance without accepting patient, phenotype, family, segregation, or private case data.
Use Folklore Clinical Variant Interpretation MCP to retrieve structured public variant evidence, automated variant-level ACMG/AMP decision support, provenance, and source-linked literature for professional review. Keep the workflow limited to public identifiers and preserve every explicit outcome state. Adapter 1.5.0 also provides ClinGen Gene-Disease Validity assertions; source coverage is bounded, not every known association.
Folklore Clinical Variant Interpretation MCP is published by Helena Bioinformatics. Its hosted endpoint is:
https://api.helena.bio/folklore/v1/mcp
No account or API key is required. The public Apache-2.0 adapter and contract are available at https://github.com/helena-bioinformatics/folklore-mcp.
A host without native MCP support can make the same public JSON-RPC call:
curl --silent --show-error --fail-with-body --max-time 60 \
-X POST https://api.helena.bio/folklore/v1/mcp \
-H 'Content-Type: application/json' \
-H 'Accept: application/json, text/event-stream' \
-H 'MCP-Protocol-Version: 2026-07-28' \
-H 'Mcp-Method: tools/call' \
-H 'Mcp-Name: search_variant_evidence' \
-d '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"_meta":{"io.modelcontextprotocol/protocolVersion":"2026-07-28","io.modelcontextprotocol/clientCapabilities":{}},"name":"search_variant_evidence","arguments":{"assembly":"GRCh38","query":"rs80357914"}}}'
The routing headers must match the JSON-RPC method and tool name. A successful
HTTP response is not sufficient: inspect JSON-RPC error, then
result.structuredContent.adapter_error, then this tool's
result.structuredContent.result.status. Other tools use different response
shapes; use the contract table.
Inspect the returned outcome before continuing. This example can return
ambiguous with multiple candidates: stop and request an unambiguous public
variant notation instead of selecting a candidate automatically.
Use this skill when the task is one public variant to structured Folklore evidence, explicit resolution-state handling, variant-linked literature, or ClinGen gene-to-disease/disease-to-gene assertions.
database-lookup for broad direct queries across ClinVar, dbSNP, gnomAD,
Ensembl VEP, COSMIC, or multiple databases.genomic-coordinates first when the assembly, coordinate convention,
contig name, or variant representation is uncertain.Folklore Clinical Variant Interpretation MCP complements those skills with one source-linked public evidence contract. It does not replace direct database review or qualified clinical judgment.
Before a variant tool call:
unsupported status exists.Accepted public variant forms include genomic coordinates, genomic/coding/
protein HGVS, SPDI, rsID, or a canonical_key returned by Folklore Clinical
Variant Interpretation MCP.
Connect to the hosted endpoint and call tools/list. Verify the available tools
instead of relying on model memory. The documented public catalog contains:
search_variant_evidencesearch_variant_literatureget_publication_detailssearch_literature_corpusget_gene_disease_associationssearch_disease_genesThe separate seventh tool support_helena is not scientific evidence; use it only when explicitly requested.
If discovery or a tool call fails, preserve the failure as an availability problem. Do not reinterpret it as lack of scientific evidence.
Read the public MCP contract before composing tool calls or interpreting response states.
Use get_gene_disease_associations for one exact gene symbol or HGNC identifier, or search_disease_genes for an exact MONDO identifier or public disease-name substring. Both accept limit (default 20, 1–50) and offset (default 0, 0–1000). See the reference for request examples. This is a separate source lookup and requires no variant input or assembly.
Preserve each returned disease identity, inheritance, evidence assessment, source
URL, date, source.version and source.snapshotSha256. This is a local ClinGen
snapshot, not a guarantee of the latest ClinGen release. Follow
pagination.nextOffset while it is present; preserve pagination-ceiling warnings.
An empty page after the total has been passed does not mean the initial query had
no matches. Do not combine distinct diseases or silently choose among name matches.
Gene-disease validity does not classify a particular variant. A not_found
response means no matching assertion in the available source, not no association.
No patient, phenotype, family, segregation, private case data or sequencing files
may be sent. Qualified professional review remains required.
Call search_variant_evidence with:
assembly: GRCh38
query: <one public variant identifier or notation>
Do not add phenotype, disease, patient, family, or treatment context to this call. Preserve the returned contract fields, source links, limitations, and usage boundary.
Read transport/JSON-RPC errors and adapter_error first. For
search_variant_evidence, let envelope = result.structuredContent; only if
envelope.result is non-null, branch on envelope.result.status. An
invalid_arguments adapter error is distinct from scientific invalid_request.
Preserve isError and any typed failure, including resolution_unavailable.
Treat the status as a control-flow value, not prose:
| Status | Required action |
|---|---|
resolved | Reuse the returned canonical_key; review the structured interpretation, provenance, source links, and limitations. |
ambiguous | Show the returned candidates and ask for an explicit public variant selection. Never choose a candidate automatically. |
not_found | Report that no result was found within this service and query scope. Do not claim universal absence. |
invalid_request | Report the validation problem and request a corrected public variant. Do not silently reinterpret the input. |
unsupported | State the relevant service boundary and stop. Do not force the query into a supported form. |
resolution_unavailable | Report a temporary resolution or availability failure. Do not treat it as evidence absence. |
Only a resolved result may proceed automatically into a variant-linked
literature workflow. Its canonical key is envelope.result.identity.canonical_key.
If envelope.result.interpretation.status is unavailable, preserve its typed
error; identity resolution has succeeded, but classification has not. Do not read
or invent a classification for that outcome.
For a resolved result:
canonical_key.After a resolved evidence call, pass the returned canonical_key to
search_variant_literature. Keep assembly as GRCh38. An optional question
may narrow the literature focus, but it must remain a public scientific question
and must not contain patient context.
Distinguish each result's match type:
exact_variant: direct match to the resolved variantvariant_alias: match through a reported aliasgene_association: broader gene-level association, not variant-specific proofLiterature associations do not alter the returned ACMG/AMP classification.
Call get_publication_details only with a PMID returned by the literature tools.
Preserve PubMed URLs, DOI/PMCID fields when present, retraction status, and the
distinction between gene mentions and variant mentions.
Use search_literature_corpus for a public natural-language scientific question
or for discovery by publication identifier, gene, variant, phenotype, HPO, or
OMIM concept. Treat results as source-linked candidates for professional review.
A zero-result response means no result was returned for that bounded query, not
that no relevant publication exists anywhere.
The query is 3–200 characters, with limit 1–25 and sort set to relevance,
newest, or oldest. For another page, reuse the returned opaque next_cursor
with the same query and sort; do not construct an offset. Keep match_types and
article_entities distinct from variant-literature match types. Report
semantic_index_used and semantic_degraded_reason; a returned lexical match
does not prove semantic retrieval worked. Preserve additional retrieval metadata
returned by the live service.
Do not place patient information into a corpus query, even if the query is not variant-specific.
Include:
GRCh38 assembly.canonical_key.This is public, variant-level decision support for qualified professional review. It does not evaluate patient, phenotype, family, segregation, or private case data and is not a diagnosis or treatment recommendation.
Use the public rsID rs80357914 to test ambiguity handling:
Call search_variant_evidence with assembly GRCh38 and query rs80357914. If the
result is ambiguous, list the returned candidates and stop for explicit
selection. Do not select a candidate or call downstream literature tools.
The ambiguity branch passes only if an ambiguous response causes the workflow to stop without automatic candidate selection. If the live response changes, record the actual status; a resolved response does not test ambiguity handling.
On 2026-09-30, live discovery reported adapter 1.5.0 and protocol 2026-07-28.
The rsID example returned ambiguous; a separate public HGVS query resolved and
was chained through its returned key to literature and a returned PMID. Corpus
cursor pagination and gene-disease offset pagination were exercised. These are
dated protocol checks, not validation of clinical accuracy. See the reference
for exact response shapes and source/live differences.
io.github.helena-bioinformatics/folklore