mif-docs
One skill per document genre. Every doc, two readers — written once, read by a person and a parser.
These docs follow the Diátaxis framework — four modes, each
matched to a different need — plus the architecture, the decision records, and the
operational runbooks. Every page is itself authored with the suite’s skills and
validated by mif-validate: the same artifact a person reads and a parser resolves.
New here? Start with the tutorialLearning-oriented. Install the suite, then author and validate your first MIF document step by step.
Skill catalog by purposeInformation-oriented. Every skill in the suite — orchestrator, authoring helpers, the four Diátaxis quadrants, architecture, decisions, specs, operations, and changelog — each with detailed exposure.
Genre & CLI quick referenceInformation-oriented. The terse lookup: genres, doc-set recipes, and every script in scripts/ with its arguments and exit behavior.
How-to guidesTask-oriented. The recipe for authoring, validating, and converting a document between Markdown and JSON-LD.
ExplanationUnderstanding-oriented. Why each document is at once a human artifact and a machine-conformant unit.
Witness your documents' provenanceTask-oriented. Turn on hook-observed provenance, watch it appear, and know what to do when a document looks wrong — opt-in, and your 'no' always wins.
ArchitectureThe suite's own arc42 and C4 architecture — how the engine, genres, and MIF substrate fit together.
Decisions (ADRs)The architecture decision records — the context and consequences behind each choice.
MIF ecosystem docsThe broader MIF ecosystem documentation, served at the org root. The normative spec lives at mif-spec.dev.