keeler
CLAUDE.md4 starsChanged 2 months agoMITView on GitHub
The Keeler workflow rules live in their own file — the plugin's, at the root of this repository, which is the plugin. In an adopting project the SessionStart hook prints it; here the import below does the same job: Those rules are the law, and two chapters sit beside them in the plugin root: graph-mode.md and gates.md. docs/KEELER.md says why they are shaped that way, and README.md is the front door — install, first day, and what Keeler is not.…
In this repository2 other files it gives its agents.
- gherkin-specsSkill
- property-testingSkill
The file
We split this CLAUDE.md 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
- Distinctive: line 65
A comment that restates the code is not allowed: Delete it. A comment earns its place by saying what the code cannot: why this way and not …
≈ “Common” is an estimate: at least 1 in 5 of the 27,641 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.
Files
CLAUDE.md
This file provides guidance to Claude Code when working with code in this repository.
The Keeler workflow rules live in their own file — the plugin's, at the
root of this repository, which is the plugin. In an adopting project the
SessionStart hook prints it; here the import below does the same job:
@keeler.md
Those rules are the law, and two chapters sit beside them in the plugin root: graph-mode.md and gates.md. docs/KEELER.md says why they are shaped that way, and README.md is the front door — install, first day, and what Keeler is not. What follows here is this repository only.
<!-- Add project-specific instructions below this line. -->How this repository diverges from the rules it ships
keeler.md is the deliverable: it is written for the projects Keeler is
installed into, and says nothing about Keeler itself. The places
where this repository cannot follow its own rules literally are recorded
here instead — project-specific instructions, in the file meant for them.
- The deliverable is shell, not Rust. The crate is the installer's
test harness, so the suite is not one
tests/acceptance.rsbut per-area files:tests/installer.rs(spec 01),tests/release.rs(spec 02) andtests/wild.rs(spec 03). One test per scenario, named after it, as usual. - Proptest seeds live beside the tests. With no
lib.rsto anchor proptest's default path, the harness setsFileFailurePersistence::WithSource("proptest-regressions"), so seeds land attests/installer.proptest-regressions. Commit them. - The gates measure
xtask/andkeeler-top/. The harness itself has no library or binary target — its job is to driveinstall.shas a subprocess — so coverage, CRAP and mutation see the two members that have one, and nothing else. Both are real code doing a real job, socrap-baseline.jsonis committed andjust crap-deltaratchets against it. Until spec 04 there was nothing here to measure at all, and the recipes said so rather than failing (spec 01). The two members are not alike:xtask/is machinery that never reaches an adopter, whilekeeler-top/ships inside the plugin and is built from its tree bykeeler keeler-top. Its own suite iskeeler-top/tests/top.rs, beside the crate rather than in the root harness, because that is the only placeCARGO_BIN_EXE_keeler-topresolves; the recipes that launch it are shell, so they are tested from the harness like every other recipe (spec 10). - Run the gates with the workspace selected. The shipped recipes
already pass
--workspace; a barecargo nextest runorcargo mutantsin this repository silently skips the xtask member and reports success having tested nothing.
The demo board
python3 demo/board.py builds a throwaway repository whose board shows
every state at once, so keeler-top can be filmed or looked at after a
change to the frame without spending an agent on a wave;
--clean takes it away again. It is repository machinery — never
installed, never reached by a recipe — and demo/README.md says what it
fabricates and what is real.
Comments
A comment that restates the code is not allowed. Delete it. A comment
earns its place by saying what the code cannot: why this way and not the
obvious one, what breaks if it changes, which upstream behavior it is
matching. # increment the counter above n += 1 is noise; the tests'
Given/When/Then lines and a rationale a reader could not derive from the
code are not.
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
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.

