okf-cli-0.5.0.0: help/okf.md
OPEN KNOWLEDGE FORMAT (OKF)
OKF is an open, human- and agent-friendly format for representing knowledge:
the metadata, context, and curated insight that surrounds data and systems.
It is meant to be authored by people, generated by agents, exchanged across
organizations, and consumed by both.
The format is intentionally minimal -- a directory of Markdown files with YAML
frontmatter. There is no schema registry, no central authority, and no
required tooling. If you can "cat" a file you can read OKF; if you can
"git clone" a repo you can ship it.
WHY IT LOOKS THIS WAY
OKF standardizes only the small set of conventions needed to make a
knowledge corpus self-describing. It aims to be:
Readable by humans, without tooling.
Parseable by agents, without bespoke SDKs.
Diffable in version control.
Portable across tools, organizations, and time.
CORE IDEAS
Bundle A self-contained directory tree of knowledge documents -- the
unit of distribution (a git repo, a tarball, or a subdirectory).
Concept A single unit of knowledge, written as one Markdown document. It
may describe a tangible asset (a table, an API) or an abstract
idea (a metric, a business process).
Concept ID The concept file's bundle-relative path with the .md suffix
removed, e.g. tables/orders.md -> tables/orders.
A concept document is YAML frontmatter (metadata) followed by a Markdown body
(free-form knowledge). Only the frontmatter "type" field is required:
---
type: PostgreSQL Table
title: Orders
description: One row per completed customer order.
---
# Orders
Orders join to [Customers](/tables/customers.md).
WHAT THE okf TOOL DOES
validate Check frontmatter conformance and, as an authoring aid, that
links resolve and log.md files are well structured.
index Generate progressive-disclosure index.md files per directory.
log Preview log.md files, check stale concepts, append entries.
graph Extract a JSON node/edge graph from links between concepts.
show Print one concept's metadata and body.
trust Report each concept's trust tier, status, and staleness.
sources List the provenance each concept records.
computations List the attested computations a bundle declares.
id Allocate and list profile-declared document IDs.
profile List and inspect profiles published by a registry.
config Show and manage okf configuration.
kit Install and manage agent skills and subagents.
assist Launch an interactive agent session with installed okf skills.
completions Generate a shell completion script.
STANDALONE BY DESIGN
The okf CLI works on plain files only. It does not require Mori, Mina, an
LLM, PostgreSQL, or network access. It tracks Google's Open Knowledge Format
v0.2 specification (the knowledge-catalog okf SPEC.md).
It reads a v0.1 bundle unchanged. A bundle that declares nothing is read
with v0.1 fallbacks and is never penalized for it; declaring
okf_version: "0.2" in the root index.md opts into the stricter reading.
SEE ALSO
okf help format Bundle layout, concept IDs, frontmatter, and links.
okf help validation What "conformant" means and how the tool checks it.
okf help profiles Checking a bundle against house conventions.
okf help index Generated index.md files and the version declaration.
okf help log log.md upkeep and the two staleness checks.
okf help graph The concept graph and what becomes an edge.
okf help trust Trust tiers, staleness, and recorded provenance.
okf help computations Listing and printing attested computations.
okf help ids Profile-declared document IDs such as ADR-7.
okf help interactive Picking a bundle and concept with fzf.
okf help config Config files, defaults, and agent settings.
okf help kit Installing and publishing agent skills and subagents.
okf help agents Installing agent skills and launching assist.
The full user guide lives under docs/user/ in the okf repository.