agentleFS
jdiejimjdiejim/chiltepin/CLAUDE.md

chiltepin

CLAUDE.md4 starsChanged 49 days agoMITView on GitHub

This repo is Chiltepin: documentation-as-code where a doc is Markdown with typed, fenced YAML blocks, and the files on disk are the single source of truth. When creating or changing any document under docs/*/.md, follow the authoring skill at skills/chiltepin/SKILL.md. It defines the block grammar and the doc#id reference scheme. The field contract for any block comes from the CLI, generated from the schema: run node packages/cli/dist/bin.js block <type> (or chiltepin block <type>) before writing a block; never look for…

We split this CLAUDE.md into the rules it gives an agent

6 rules. About 0 are common, 2 are distinctive, and 4 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 4

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    When creating or changing any document under docs//*.md, follow the authoring skill at skills/chiltepin/SKILL.md.

    line 10
  2. Distinctive:

    Use only the documented block types (107 of them, plus 12 permanent aliases for merged old names) and their documented fields — the schemas …

    line 20
  3. Very specific:

    Read ARCHITECTURE.md before changing structure.

    line 30
  4. Very specific:

    Keep pnpm lint, pnpm typecheck, and pnpm test green.

    line 39
  5. Distinctive:

    No I/O, no DOM. Everything depends inward on core.

    line 52
  6. Very specific:

    Never claim Chiltepin output "is STE" or "is STE-compliant".

    line 96

≈ “Common” is an estimate: at least 1 in 5 of the 25,680 CLAUDE.md 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
CLAUDE.md
<!-- Repo location: CLAUDE.md at the repository root. Claude Code reads this automatically. -->

Chiltepin — agent guide

This repo is Chiltepin: documentation-as-code where a doc is Markdown with typed, fenced YAML blocks, and the files on disk are the single source of truth.

Writing or editing documentation

When creating or changing any document under docs/**/*.md, follow the authoring skill at skills/chiltepin/SKILL.md. It defines the block grammar and the doc#id reference scheme. The field contract for any block comes from the CLI, generated from the schema: run node packages/cli/dist/bin.js block <type> (or chiltepin block <type>) before writing a block; never look for a hand-written contract file. In short:

  • Prose is plain Markdown; structure goes in typed blocks (e.g. sequence, erd, table, callout, c4, flow, timeline, userstory). Never paste raw HTML or inline SVG.
  • Use only the documented block types (107 of them, plus 12 permanent aliases for merged old names) and their documented fields — the schemas are strict. Bodies are YAML.
  • Give a block an id: when it needs to be referenced; reference it as doc#id.
  • Edit the specific block surgically — don't regenerate whole files.
  • Run chiltepin check and fix all diagnostics before finishing. A change isn't done until it passes.

Working on the codebase itself

  • Read ARCHITECTURE.md before changing structure.
  • Dependency direction points inward to chiltepin-core; core is pure (no I/O, no DOM).
  • Block types are defined once in the block registry; adding one means adding a schema in core plus a renderer in each target. Registries are compile-time exhaustive — don't bypass them with ad-hoc switches.
  • Libraries return diagnostics/typed results; only the CLI throws and sets exit codes.
  • chiltepin-studio BUNDLES core/render (devDependencies, baked in by Vite at publish). A changeset that changes anything Studio renders must also patch chiltepin-studio, or the published Studio canvas keeps the old renderer.
  • Keep pnpm lint, pnpm typecheck, and pnpm test green. Conventional commits.

Chiltepin — what this repo is

Documentation-as-code. A doc is plain Markdown; every visual thing is a typed block — a fenced section with a type and a YAML body. 107 block types across 13 families (narrative, tables/code, API, architecture, flows/state, data model, charts, planning, business/decisions, design system, algorithms, AI/agents). .md files on disk are the only source of truth. Nothing else holds state.

Packages

  • chiltepin-core — pure library. Parsing, block registry, schemas, diagnostics. No I/O, no DOM. Everything depends inward on core. If a change here is not additive, it is a breaking change to every other package — say so.
  • chiltepin-render — deterministic renderers. One block type → one renderer → consistent HTML/SVG. The LLM never draws. It writes ~1KB of YAML; the renderer owns 100% of layout and geometry.
  • chiltepin (CLI, chiltepin binary) — validation, static sites, HTML, slides, PDF, Studio, block references, and the built-in demo. Run chiltepin --help for commands.
  • chiltepin-studio — web canvas/editor bundling core + render.
  • Authoring skill (skills/chiltepin/SKILL.md + reference files, one copy in the repo) — teaches an agent to author docs. It installs into any agent with npx skills add jdiejim/chiltepin -y; the CLI build copies it into the package for chiltepin skill.

The invariant that matters most

Geometry is code, never prompt. If a task tempts you to teach the model pixel coordinates, inline SVG, or layout rules, you have misdiagnosed the task: the fix belongs in chiltepin-render, not in the skill. The skill teaches selection, composition, and content. The renderer owns appearance.

Definition of done

pnpm -w typecheck && pnpm -w test && pnpm -w lint && chiltepin check all green, plus a rendered example of anything visual you touched. Schema changes ship with: the schema, the renderer, a reference entry in the skill, a catalog example, and a test. Five things. A block type missing any of them does not exist.

House style for anything a user reads

Short, technical, synthesized. No filler openers ("In this section we will…"), no restating the block below in prose, no adjective stacking. Explain to a smart non-specialist: simple words, precise claims. If a sentence carries no fact, delete it.

Technical text in this repo follows STE discipline — the writing rules of ASD-STE100 Simplified Technical English, adapted for software docs. The rules live in skills/chiltepin/reference/style-ste.md. Read that file before you write any prose block, CLI help string, diagnostic message, or skill instruction.

Two hard constraints on how we use it:

  • ASD-STE100 is free to obtain but not free to redistribute. Never copy the specification text or its ~900-word approved dictionary into this repo. We apply the rules and keep our own term list.
  • Never claim Chiltepin output "is STE" or "is STE-compliant". The wording is "STE-informed" or "follows STE writing discipline". Certified compliance requires the real dictionary, which we do not ship.

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.