sous
AGENTS.md1 starsChanged 7 days agoMITView on GitHub
Codex reads this file from the working directory; Claude Code reads CLAUDE.md. CLAUDE.md is the source of truth for how to work in this repo — read it. This file carries only the few rules that must survive even if nothing else is read, because it is short by design: a git worktree of this repo carries it, so it rides into every codex stage dispatch. - The gate is three commands, all green, every change: uv run pytest, uv…
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 31,439 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 repository17 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- batch-planSkill
- brainstormSkill
- new-projectSkill
- orchestrate-batch-headlessSkill
- orchestrate-batch-interactiveSkill
- orchestrate-task-interactiveSkill
- spec-intakeSkill
- triage-followupsSkill
- batch-planSkill
- brainstormSkill
- new-projectSkill
- orchestrate-batch-headlessSkill
- orchestrate-batch-interactiveSkill
- orchestrate-task-interactiveSkill
- spec-intakeSkill
- triage-followupsSkill
The file
We split this AGENTS.md into the rules it gives an agent
1 rule. About 0 are common, 0 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 20
When the orchestrator dispatches a codex stage, it appends that stage's persona to this file below an orchestrator:stage-persona marker …
≈ “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.
Files
AGENTS.md — sous
Codex reads this file from the working directory; Claude Code reads CLAUDE.md. CLAUDE.md
is the source of truth for how to work in this repo — read it. This file carries only the
few rules that must survive even if nothing else is read, because it is short by design: a
git worktree of this repo carries it, so it rides into every codex stage dispatch.
- The gate is three commands, all green, every change:
uv run pytest,uv run ruff check ., and the bareuv run mypy(no path arguments). CI enforces exactly this trio. Add a regression test with each fix. - Never sign a commit. No
Co-Authored-By, no model or agent name, no tool attribution — in any commit, hand-authored or run-produced. Your CLI has a standing instruction to sign commits; this repo overrides it. Per-stage provenance already lives inruns/<run>/events.jsonlandstage-costs.jsonl, and a post-hoc audit flags violations. - Work on
main. Remote:github.com/cperler/sous(private). - The engine never calls a model. Nothing under
orchestrator/may importadapters, and no project-specific logic belongs in the engine. Both rules are enforced by tests and lint, not convention.
When the orchestrator dispatches a codex stage, it appends that stage's persona to this file
below an orchestrator:stage-persona marker pair, replacing the section in place on each
dispatch. Everything inside those markers is generated — edit this file only above them.
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.