agentleFS
Sign inStart free
erickdronskierickdronski/gtm-skills/AGENTS.md

gtm-skills

AGENTS.md0 starsChanged 27 days agoMITView on GitHub

Instructions for coding agents. Humans should read CONTRIBUTING.md. A pack of go-to-market skills sitting on gtmkit, a dependency-free Python library that does the arithmetic. The premise is that numbers in business documents should come from tested code rather than from a language model, so the tests are the point, not overhead. - Standard library only. CI fails if requirements.txt appears. - Python 3.9 compatible. No X | Y unions, no match. - Never use eval or exec. Formula evaluation goes…

  • 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 31,774 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 repository9 other files it gives its agents.

The file

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

2 rules. About 0 are common, 1 is distinctive, and 1 is 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.

  • 1

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    If you change a number that appears in README.md, regenerate it and update the README.

    line 34
  2. Very specific:

    Keep the body under 500 lines.

    line 54

≈ “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

Working in this repository

Instructions for coding agents. Humans should read CONTRIBUTING.md.

What this is

A pack of go-to-market skills sitting on gtmkit, a dependency-free Python library that does the arithmetic. The premise is that numbers in business documents should come from tested code rather than from a language model, so the tests are the point, not overhead.

Non-negotiables

  • Standard library only. CI fails if requirements.txt appears.
  • Python 3.9 compatible. No X | Y unions, no match.
  • Never use eval or exec. Formula evaluation goes through gtmkit/expr.py, which walks an AST against a whitelist. Specs come from untrusted sources.
  • Return None, never a fabricated number. A missing IRR is information.
  • Report what you drop. Filtered rows, rejected inputs, and truncated lists must appear in the output.

Before you finish

python3 -m unittest discover -s tests -t .
python3 tools/validate_skills.py

Both must pass. The linter checks skill frontmatter, description trigger quality, dead links, referenced gtmkit modules and the flags their commands use, and length budgets.

If you change a number that appears in README.md, regenerate it and update the README. A README quoting output the code no longer produces is worse than no README.

Layout

gtmkit/          the engine — one module per analysis, each a CLI and a library
skills/          one directory per skill: SKILL.md, references/, assets/
tests/           unittest, no pytest, no plugins
tools/           repository tooling (skill linter)
examples/        every example is exercised by CI

Writing skills

SKILL.md frontmatter needs name (matching the directory) and description. The description is the only thing that determines whether a skill triggers — it must state when to use the skill and name phrasings a real user would type.

Keep the body under 500 lines; move detail to references/. Include failure modes, and explain the reasoning rather than issuing rules — the instructions are read by a capable model that does better with a stated mechanism than with an unexplained MUST.

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