agentleFS
arto-apparto-app/Arto/AGENTS.md

Arto

AGENTS.md261 starsChanged 3 days agoMITView on GitHub

- Test first. Red → green → refactor. Write the failing test before the code that makes it pass. - Each layer holds one thing. Put a fact in the one layer that owns it, and nowhere else. Before writing, fixing or reviewing a record, read docs/adr/README.md; before writing, fixing or reviewing a spec, read docs/spec/README.md. Follow their writing order. - Dependencies are current. Before adding or bumping one, look up its latest release in the registry. Do not copy…

  • Where files are headingNew AGENTS.md files have more rules about keeping secrets out of the code, and fewer about code style.

    Of the AGENTS.md files we read, the 33,315 first seen in the last six months, against the older ones: the kind of rule that gained most share, and the kind that lost most. About files like this in general, not this one, and an estimate, since a kind is a pattern in the words.

  • In this repositoryOne other file it gives its agents.

The file

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

8 rules. About 0 are common, 5 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈5

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Each layer holds one thing: Put a fact in the one layer that owns it, and nowhere else.

    line 11
  2. Very specific:

    Before writing, fixing or reviewing a record, read docs/adr/README.md.

    line 24
  3. Distinctive:

    Follow their writing order.

    line 26
  4. Distinctive:

    Dependencies are current: Before adding or bumping one, look up its latest release in the registry.

    line 27
  5. Very specific:

    Icons: use the add-icon skill

    line 35
  6. Distinctive:

    Application launch: do NOT launch the app; the user handles this

    line 36
  7. Very specific:

    Run just ci before reporting a change as done.

    line 40
  8. Distinctive:

    Add a row when a new area appears.

    line 129

≈ “Common” is an estimate: at least 1 in 5 of the 42,283 AGENTS.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
AGENTS.md

Arto

<!-- Quoted from README.md; do not rephrase it here. -->

Arto — the Art of Reading Markdown. A desktop app for reading Markdown, locally and offline.

How it is built

  • Test first. Red → green → refactor. Write the failing test before the code that makes it pass.
  • Each layer holds one thing. Put a fact in the one layer that owns it, and nowhere else.

    LayerHolds
    docs/concept.mdwhat this is and for whom
    docs/adr/one decision and why it was made
    docs/spec/the behaviour users and other components can rely on
    Teststhe details that pin the behaviour down: rules, boundaries, tables
    Codehow it is done
    Commitswhy a change was made
    Commentswhy it is not done another way

    Before writing, fixing or reviewing a record, read docs/adr/README.md; before writing, fixing or reviewing a spec, read docs/spec/README.md. Follow their writing order.

  • Dependencies are current. Before adding or bumping one, look up its latest release in the registry. Do not copy a version from memory or from another repository.

Project rules

  • Test code: indoc for multi-line strings, tempfile for files
  • Module system: Rust 2018+ (no mod.rs)
  • Icons: use the add-icon skill
  • Application launch: do NOT launch the app; the user handles this

Quality gate

Run just ci before reporting a change as done. It checks formatting (rustfmt, oxfmt, nixfmt), lints (clippy, oxlint), runs every test, and checks the rustdoc, feature combinations, advisories, unused dependencies, the crate packages, the documentation's links and the workflow files. CI runs the same recipes.

just fmt formats everything just ci checks.

Layout

PathWhat lives there
crates/arto/The desktop app (Dioxus). Windows, the panel and its faces, menus, IPC client/server
crates/arto-markdown/Markdown → HTML pipeline shared by the app, arto page and Quick Look
crates/arto-config/config.json types and file I/O (no globals)
crates/arto-keybindings/Keybinding parsing and defaults
crates/arto-lsp/The wire protocol: JSON-RPC 2.0 in LSP framing, and the socket it runs on
crates/arto-mcp/arto mcp: an MCP server over the wire protocol, for agents
crates/arto-frontend/The frontend's bundles as Rust strings: the library bundles arto-page embeds, and the runtime and stylesheet for another app's page
crates/arto-page/Standalone page renderer: arto page CLI and the Quick Look static library
crates/arto-types/Values config.json and the wire protocol both carry (Theme, FileOpenBehavior)
frontend/TypeScript and CSS bundled into each crate's assets/frontend/ by Vite
samples/Rendering samples; their HTML is snapshot-tested in arto-markdown
schemas/JSON Schema of config.json, generated from arto-config and kept in step by its tests
platform/OS-specific packaging (macOS bundle and Quick Look extension)
docs/concept, ADRs, and specs — the pages the README links to are the specs
nix/Wrappers the Nix build puts in front of macOS tools
.claude-plugin/, plugins/The Claude Code plugin marketplace: the arto plugin registers arto mcp and a skill for its tools

Where the details live

Architecture notes load on demand from .claude/rules/ when a matching file is read or edited (see each rule's paths frontmatter): windows and state, configuration, IPC, menus, the Markdown pipeline, Dioxus async patterns, testing, UI design, interface translation, ADRs and specs. .claude/TIPS.md collects longer-form lessons; read it when a rule points there.

The project description lives in README.md; quote it instead of inventing descriptions.

Language

  • README, docs and commit messages: English
  • Code comments, log and error messages: English

Commits

Scoped Commits (.gitmessage is the template):

<scope>: <description>

<why this change is needed>
  • The scope names where the change is, from the table below. Several scopes are joined with , ; a change across the whole tree uses treewide.
  • The description starts in lower case, has no trailing period, and stays within 72 characters.
  • The body explains why. The code shows how; the tests show what.
  • A breaking change says so in the body and carries a BREAKING CHANGE: trailer.

Commits before the switch use Conventional Commits (feat(sidebar): …); they stay as they are.

ScopeCovers
artocrates/arto/
markdowncrates/arto-markdown/
configcrates/arto-config/
keybindingscrates/arto-keybindings/
lspcrates/arto-lsp/
mcpcrates/arto-mcp/
pagecrates/arto-page/
typescrates/arto-types/
frontendfrontend/, crates/arto-frontend/
samplessamples/
schemasschemas/
platformplatform/
docsREADME.md, CONTRIBUTING.md, docs/concept.md, docs/images/
adrdocs/adr/
specdocs/spec/
agentsAGENTS.md, .claude/
buildjustfiles, the workspace Cargo.toml, rust-toolchain.toml, .cargo/, lint and deny settings
ci.github/
nixflake.nix, flake.lock, nix/
plugins.claude-plugin/, plugins/
depsdependency updates

Add a row when a new area appears.

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.