agentleFS
modeled-information-formatmodeled-information-format/mif-docs-plugin/skills/mif-frontmatter/SKILL.md

mif-frontmatter

Skill1 starsChanged 35 days agoMITView on GitHub

Author MIF (Modeled Information Format) Level 1-3 frontmatter for any document, climbing from the L1 floor to L2/L3 only when the drafting context supplies real detail. Use whenever a doc needs MIF-conformant YAML frontmatter that projects losslessly to the canonical JSON-LD.

The file

We split this skill into the rules it gives an agent

1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    every other field passes through verbatim (it must be a valid MIF property or an open extension key).

    line 57

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

mif-frontmatter

Shared substrate for every genre skill in this suite. It governs the YAML frontmatter that makes a document a MIF unit: a block that projects to the canonical JSON-LD and validates fail-closed against https://mif-spec.dev/schema/.

The "one artifact, two readers" contract

A MIF doc exists in two interconvertible forms — human-readable markdown (frontmatter + body) and machine-readable JSON-LD (the schema-checked canonical object). Authoring is bidirectional and the round-trip is lossless; see the mif-validate skill for the tooling.

Whenever a level, floor, or the type field itself comes up — deciding what level a document can claim, explaining the L1 floor, or drafting frontmatter — state the type enum explicitly (see the L1 row below). It's a hard constraint on an L1 field, not an implementation detail to leave implicit, so surface it every time L1 is discussed, not only when actually writing the YAML. For what distinguishes the three values and how to choose one before drafting, see Documentation Taxonomy — the canonical source for this enum; don't restate its definitions here.

Level floors (attempt the highest the context supports)

LevelFrontmatter fieldsEmit when
L1 (hard floor)id (a UUID), type (enum: semantic | episodic | procedural), created (+ body becomes content)Always. Below L1 is a skill error.
L2namespace, modified, temporalreview cadence / topic namespace known
L3provenance, citations[], relationships[]doc sourced from real, attributable input

Field projection (markdown frontmatter -> JSON-LD)

  • id -> @id (auto-prefixed urn:mif: if not already a URN). MIF 1.4.0+ requires @id to be urn:mif:<uuid>, so id MUST be a UUID: a slug such as prd-saved-carts fails the schema. Mint a random UUIDv4 for a new document; to derive a stable id from an existing slug, use UUIDv5 in the MIF namespace, uuid5(uuid5(NAMESPACE_URL, "https://mif-spec.dev"), slug), the same derivation MIF's migration script and mif-rs use. A relationship target naming another concept by URN uses that concept's urn:mif:<uuid>; a bundle-relative path (/semantic/policy.md) is the RECOMMENDED target for a concept in the same bundle.
  • type -> conceptType; MUST be one of semantic, episodic, procedural (see Documentation Taxonomy for what each means and how to pick one). @type is always Concept. The document genre (adr, tutorial, runbook…) lives in namespace/tags, not in type.
  • created -> created (ISO-8601 date-time)
  • body -> content
  • every other field passes through verbatim (it must be a valid MIF property or an open extension key).

Grade-down rule (never fabricate)

Attempt the highest level the drafting context supports, field by field. If a level's required field cannot be populated from real input (no citations -> no L3; no review cadence -> omit temporal), drop to the next lower level rather than writing a placeholder. Never emit an empty or unresolved-placeholder MIF field. This helper proposes frontmatter; mif-validate disposes.

Minimal L1 example

---
id: 7b3c1e90-5a2f-4c8d-9e10-2f6a4b8c1d3e
type: semantic
created: 2026-06-29T10:30:00Z
---

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.