Annotate Markdown documents with MIRA research objects and source-local relations.
npx skills add https://github.com/stencila/stencila --skill mira-annotation
Please help me install this Agent Skill. Skill: mira-annotation Source: https://github.com/stencila/stencila/tree/main/.stencila/skills/mira-annotation Command: npx skills add https://github.com/stencila/stencila --skill mira-annotation
Scholarly Markdown documents often contain claims, evidence, questions, protocols, and requests that are not machine-readable, making it hard to trace which evidence supports which claim. This Skill adds semantic MIRA research-object annotations directly into the source Markdown without rewriting the author's wording.
Annotate the attached manuscript.qmd with MIRA claims, evidence, and their supporting relations while preserving all Quarto syntax.
High-intent search queries and answers about installing and using this skill.
Wrap claim and evidence passages in fenced blocks matching your Markdown flavor, such as ::: claim #claim-id for Stencila Markdown or ::: {.claim #id} for Quarto. Add relation attributes like supported-by pointing to the related object id.
MIRA annotation supports Stencila Markdown (.smd), Quarto (.qmd), MyST (.myst), and plain Markdown (.md). Each dialect uses its native syntax: colon fences for Stencila, fenced divs for Quarto, and directives for MyST.
Yes, Quarto annotation preserves YAML frontmatter, executable code chunks, chunk options, callouts, and cross-references. Annotations are added as fenced divs without moving or reformatting code cells.
Avoid it for general copyediting, prose rewriting, or when you only want a separate knowledge graph without modifying the source. It is designed for semantic research-object annotation grounded in the document text.
Relations use attributes like supports, supported-by, opposes, addresses, and grounds, pointing to object ids with a leading # such as #evidence-study. Direction should match existing document conventions, and only explicit or strongly local relations are added.
groundnutyAudit LaTeX and Markdown citations against bibliographies for missing entries and malformed fields.
nahisahoExtract biomedical entities and relationships from scientific text into a literature knowledge graph.
q734738781Review manuscript text for unsupported claims and missing citations.
ZpankzBridge CSV or database domain graphs with LLM-extracted entities from PDFs and text in Neo4j.
