agentleFS
kaithoughtarchitectkaithoughtarchitect/toppingout/AGENTS.md

toppingout

AGENTS.md14 starsChanged 20 days agoMITView on GitHub

For coding agents that open this repository. Humans: start with README.md. Topping Out is a method and a set of skills for running AI coding agents one ticket at a time, for hours, without drifting. A foreman session plans each job and never writes product code; a cheaper builder model does the work; inspectors (one from another model family) review it; the owner gives the word for merges, pushes and other reserved acts. Every job starts a fresh session from…

The file

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

4 rules. About 0 are common, 1 is distinctive, and 3 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.

  • 3

    Very specific

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

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

    Don't copy files by hand. In the repo that will hold the user's records, read skills/f-setup/SKILL.md from this checkout and follow it.

    line 11
  2. Very specific:

    Run the checks for what you touched, for example node skills/f-setup/install.mjs --self-test, python skills/f-board/test_render_board.py, …

    line 32
  3. Distinctive:

    Keep the house style: plain language, no em dashes, every claim backed by a file, a command or a commit.

    line 33
  4. Very specific:

    Never name the builder's model or vendor in records (RULES.md).

    line 34

≈ “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

AGENTS.md

For coding agents that open this repository. Humans: start with README.md.

What this is

Topping Out is a method and a set of skills for running AI coding agents one ticket at a time, for hours, without drifting. A foreman session plans each job and never writes product code; a cheaper builder model does the work; inspectors (one from another model family) review it; the owner gives the word for merges, pushes and other reserved acts. Every job starts a fresh session from files and ends with a record of what was asked, built, tested and found.

If the user wants to use Topping Out in their own project

Don't copy files by hand. In the repo that will hold the user's records, read skills/f-setup/SKILL.md from this checkout and follow it. It runs skills/f-setup/install.mjs, profiles the repo, and asks only what it can't find.

  • Needs git, Node 18+, Python 3.9+ and Bash (/f-setup checks each).
  • Tested on Windows 11 with Claude Code as the foreman. macOS and desktop Linux are expected to work but are not verified yet (SETUP.md, "Tested setup"). It also runs on a Linux server through Orca (SETUP.md §8b).
  • The installer copies only what a project needs (skills, agents, templates, the method). It never copies docs/, the wiki, .github/ or this file.

Where things are

PathWhat it is
README.mdwhat it is and how a job runs, from the owner's side
METHOD.mdthe process the foreman follows, step by step
RULES.mdthe acts that need the owner's word, and the exceptions
SETUP.mdinstalling, requirements, the tested setup
skills/the skills (f- for foreman), each with a SKILL.md
agents/the review lenses and helper agents the skills call
templates/work orders, reviews, ledgers, handovers
docs/the website at toppingout.dev, generated by scripts/build_site.py: edit the sources, never the pages

If you are changing this package

  • Run the checks for what you touched, for example node skills/f-setup/install.mjs --self-test, python skills/f-board/test_render_board.py, node skills/f-comply/comply.test.mjs.
  • Keep the house style: plain language, no em dashes, every claim backed by a file, a command or a commit.
  • Never name the builder's model or vendor in records (RULES.md).

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.