agentleFS
maxtrigifymaxtrigify/Otto/AGENTS.md

Otto

AGENTS.md126 starsChanged 10 days agoApache-2.0View on GitHub

- Anything a person sees is designed before it is built. Read DESIGN.md (the design system) and follow the otto-design skill in .agents/skills/otto-design/: design in Paper on Otto's real frame and tokens, show Max on Shorelight, then write SwiftUI through @Environment(\.ottoPalette). - Project skills live in .agents/skills/ and are linked from .claude/skills/. frontend-design applies to all UI; components-build and superdesign are for web surfaces only.

  • 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 31,542 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 repository5 other files it gives its agents.

The file

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

9 rules. About 0 are common, 3 are distinctive, and 6 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 6

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  • could be automatic: a setting could enforce it
  1. Very specific:

    Use one active coding task per Git worktree and one unique codex/...

    line 5
  2. Distinctive:

    Treat the repository root on main as the integration checkout.

    line 6
  3. Very specific:

    Before editing, run pwd, git branch --show-current, and git status --short.

    line 7
  4. Distinctive:

    Never remove a worktree until its dirty state and commits have been audited.

    line 9
  5. Very specific:

    Use scripts/otto-worktree remove <task-name> for cleanup.

    line 10
  6. Very specific:

    Never invoke swift build, swift test, swift run, or swift package directly during agent work.

    line 14
  7. Very specific:

    Use scripts/otto-swift <command> ....

    line 15
  8. Very specific:

    If SwiftPM appears stuck, run scripts/otto-swift-doctor.

    line 18
  9. Distinctive:

    Commit completed work on the task branch and report the worktree path, branch, and commit SHA.

    line 23

≈ “Common” is an estimate: at least 1 in 5 of the 40,246 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

Otto agent working agreement

Checkout ownership

  • Use one active coding task per Git worktree and one unique codex/... branch.
  • Treat the repository root on main as the integration checkout. Do not perform parallel feature work there.
  • Before editing, run pwd, git branch --show-current, and git status --short.
  • Do not create worktrees under /tmp, .claude, application caches, or a second workspace root. Use scripts/otto-worktree create <task-name> so durable agent worktrees live under the single sibling Otto-worktrees/ directory. Worktrees created before the rename stay in the sibling Muse-worktrees/ directory; otto-worktree remove still finds them. Never move an existing worktree.
  • Never remove a worktree until its dirty state and commits have been audited.
  • Use scripts/otto-worktree remove <task-name> for cleanup. It refuses dirty worktrees and archives the final commit before removal.

SwiftPM

  • Never invoke swift build, swift test, swift run, or swift package directly during agent work.
  • Use scripts/otto-swift <command> .... It assigns a cache and scratch directory unique to the current worktree, prevents concurrent use of that directory, applies a timeout, and cleans up the complete process group when cancelled.
  • Normal builds and tests must use Package.resolved without remote updates. Use scripts/otto-swift resolve only as a deliberate integration task and commit any resulting Package.resolved change.
  • Set OTTO_SWIFT_TIMEOUT=0 only for an intentionally unbounded interactive run.
  • If SwiftPM appears stuck, run scripts/otto-swift-doctor. Do not use broad commands such as pkill swift or killall git.
  • The old scripts/muse-swift, muse-worktree and muse-swift-doctor names are thin wrappers kept for agents that still call them. The pre-rename MUSE_SWIFT_* variables are still read when the OTTO_SWIFT_* one is unset, and SwiftPM state stays in an existing ~/Library/Caches/Muse/SwiftPM until the app migrates it to ~/Library/Caches/Otto/SwiftPM.

Handoff

  • Commit completed work on the task branch and report the worktree path, branch, and commit SHA.
  • Report uncommitted changes explicitly. Do not silently move or copy them between worktrees.
  • Integrate task commits into main from the integration checkout using a reviewed merge or cherry-pick.

Design

  • Anything a person sees is designed before it is built. Read DESIGN.md (the design system) and follow the otto-design skill in .agents/skills/otto-design/: design in Paper on Otto's real frame and tokens, show Max on Shorelight, then write SwiftUI through @Environment(\.ottoPalette).
  • Project skills live in .agents/skills/ and are linked from .claude/skills/. frontend-design applies to all UI; components-build and superdesign are for web surfaces only.

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.