agentleFS
Sign inStart free
gregfalettogregfaletto/r-package-dev-skill/AGENTS.md

r-package-dev-skill

AGENTS.md0 starsChanged 41 days agoMITView on GitHub

Entry point for harnesses that read a repo-root agent file. This file is a pointer, not a copy. SKILL.md is the real entry point: who this is for, the cycle, the non-negotiables, and the routing table for everything in references/. Read it now, before doing anything else. Then, if you are working inside an R package: 1. .workflow/PROFILE.md in that package — its build and gate commands, conventions, dependencies, and gotchas. Several of its fields switch what the process does.…

  • 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 repositoryOne other file it gives its agents.

The file

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

6 rules. About 0 are common, 1 is distinctive, and 5 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.

  • 5

    Very specific

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

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

    If you are editing *this repo* rather than using it: process content belongs in references/ so it exists once.

    line 29
  2. Very specific:

    Before you finish, run Rscript scripts/check-docs.R and Rscript scripts/check-briefs.R, plus Rscript scripts/check-fields.R if you touched …

    line 32
  3. Very specific:

    Before proposing a change to the skill, check CONSIDERED.md for it.

    line 37
  4. Distinctive:

    Commit as you go. Push, change the repo's settings, or do anything else outward-facing only when the maintainer asks.

    line 39
  5. Very specific:

    If a LOCAL.md exists at the repo root, read it.

    line 43
  6. Very specific:

    Don't keep instructions or facts about this skill in an agent's memory feature.

    line 47

≈ “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

Entry point for harnesses that read a repo-root agent file. This file is a pointer, not a copy.

Read this first

SKILL.md is the real entry point: who this is for, the cycle, the non-negotiables, and the routing table for everything in references/. Read it now, before doing anything else.

Then, if you are working inside an R package:

  1. .workflow/PROFILE.md in that package — its build and gate commands, conventions, dependencies, and gotchas. Several of its fields switch what the process does. If it is missing, SKILL.md says how to create it.
  2. references/harness-notes.md — how to spawn subagents in your harness, and why absolute paths in every brief are non-negotiable.

What this is

A development process for statistical and numerical R packages — the kind implementing an estimator, a model, or an algorithm against a documented methodology, where a wrong answer looks exactly like a right answer: nothing crashes, no type check fails, and often the test that should have caught it came from the same derivation as the bug.

Working on the skill itself

If you are editing this repo rather than using it: process content belongs in references/ so it exists once. Anything harness-specific goes in references/harness-notes.md.

Before you finish, run Rscript scripts/check-docs.R and Rscript scripts/check-briefs.R, plus Rscript scripts/check-fields.R if you touched an instruction or templates/PROFILE.md. What each list in their output obliges you to do is in CONTRIBUTING.md.

Before proposing a change to the skill, check CONSIDERED.md for it.

Commit as you go. Push, change the repo's settings, or do anything else outward-facing only when the maintainer asks. The maintainer also edits files on GitHub's website, so fetch and check git status -sb before you edit or push.

If a LOCAL.md exists at the repo root, read it. It is gitignored, and it records facts about this machine's install, such as where the skill is installed and which installed files are copies.

Don't keep instructions or facts about this skill in an agent's memory feature. What an agent needs to use the skill belongs in SKILL.md and references/, what it needs to edit the skill belongs in this file or CONTRIBUTING.md, and what is specific to one machine belongs in LOCAL.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.

Check your setup, free