togomcp

Generate or update MIE (Metadata Interoperability Exchange) YAML files that describe RDF databases for TogoMCP. Use this skill whenever the user asks to create, write, regenerate, update, or improve an MIE file for any RDF database (UniProt, ChEMBL, Reactome, Rhea, PubChem, MeSH, PDB, or any new bio/chem database being added to TogoMCP), or mentions anything about the `togo_mcp/data/mie/` directory, MIE schema documentation, database onboarding for TogoMCP, or describing an RDF endpoint's schema/ShEx/SPARQL examples for LLM consumption. Trigger even if the user says things like "document this database", "add Reactome to our MIE collection", "the ChEMBL metadata file is out of date", or "I need to describe this SPARQL endpoint for Claude" — as long as the context is TogoMCP or an RDF database being described in YAML.

Generate, refresh, or replace benchmark questions for the TogoMCP evaluation set (benchmark/questions/question_XXX.yaml). Use whenever the user asks to create, add, extend, refresh, re-validate, replace, redirect, or retire benchmark questions — e.g. "add a few benchmark questions", "extend the QA set", "we need more list-type questions", "write a new question about a topic or database", "grow the benchmark past 50", "refresh question_042", "the answer to Q030 looks stale", "the queries are out of date". Anything mentioning benchmark/questions, coverage_tracker.yaml, the QA creation guide, or question YAML files. New questions follow the v5.5.0 type-first protocol against live RDF databases (SPARQL, OLS4, PubMed); updates re-run queries against current state and refresh answers — both presented for approval before writing.

Systematically validate a research article's claims against TogoMCP RDF databases — molecular formulas, reaction equations, pathways, protein function, GO definitions — instead of trusting the paper's text or keyword-search snippets. Use whenever the user asks to "analyze this paper/article", fact-check or validate a study's biology, verify claimed metabolites/reactions/pathways/ proteins/processes against databases, score the evidence for a paper's claims, or build a cross-database evidence chain (ChEBI → Rhea → UniProt → Reactome → GO). Triggers on "article analysis", "validate this study", "check this paper's claims". The full method lives in references/research_article_analysis.md.

Run a multi-scale disease pathophysiology analysis — molecular defects through to clinical symptoms — using TogoMCP (RDF Portal SPARQL), TogoID ID conversion, OLS4, and PubMed. Use whenever the user asks to "analyze a disease", build a disease mechanism model, find disease-associated proteins/pathways/drugs, map a disease across biological scales (molecular → pathway → cellular → tissue → clinical → treatment), or convert disease/protein/drug IDs across databases for a disease. Also triggers on "disease analysis", "pathophysiology of X", "what proteins/pathways/drugs are involved in X". The full method lives in references/disease_analysis.md.

PRISM finds entities (genes, proteins, compounds) at the INTERSECTION of two or more property sets on RDF knowledge graphs via TogoMCP — e.g. "genes common to disease X and druggable pathway Y", "targets both linked to phenotype P and modulated by existing drugs". Use whenever the user asks a set-intersection / overlap / "what is common to" / "is there a gene or target that is both ... and ..." question, a drug-repurposing or target-identification question, or anything combining a DISEASE/PHENOTYPE axis with a FUNCTION axis and/or a DRUGGABILITY axis. Trigger even on "shared genes between A and B", "druggable targets for disease X", or "intersection of ... and ...". PRISM makes each axis a reproducible predicate expanded over the ontology hierarchy, triangulated across evidence sources, and intersected by stable IDs with a provenance ledger. If you catch yourself listing candidate genes from memory, STOP and use it.

Update the public TogoMCP intro/landing page at `togo_mcp/data/docs/togomcp-intro.html` whenever the catalog of databases, the catalog of MCP tools, or any other surface fact shown on the page changes. Trigger when the user adds/removes/renames a database (MIE file, `endpoints.csv` row), adds/removes/renames a tool (any `@mcp.tool` decorator in `togo_mcp/*.py`), updates the preprint citation, changes setup instructions for a host (Claude/ChatGPT/Gemini), revises the recommended companion MCP servers, or otherwise edits something the landing page advertises. Also trigger when the user says "update the intro page", "regenerate the landing page", "refresh togomcp-intro.html", "the intro is out of date", or similar — even if they don't name the file. Spec for what the page should contain is `togo_mcp/data/docs/make_intro.md`; the rendered page is `togo_mcp/data/docs/togomcp-intro.html`.