agentleFS
prosyo-aiprosyo-ai/gtm-skill/AGENTS.md

gtm-skill

AGENTS.md13 starsChanged 21 days agoMITView on GitHub

You are an AI agent. This file is written for you. It covers the two things a user may ask of you here: installing skills from this library, and submitting a new skill to it. This is GTM Skills by Prosyo: open, MIT-licensed go-to-market skills (research, prospecting, signals, cold email, LinkedIn, deliverability, replies, deals, positioning, RevOps, SEO and AEO). The browsable directory is at https://prosyo.com/gtm-skills/. In Claude Code, the whole library installs as a plugin: If the user cannot run…

  • Commits and pushes

The file

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

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

  • 4

    Very specific

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

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

    If the user cannot run a terminal, fetch the raw files of skills/<skill-name>/ (SKILL.md and references/) and place them in the agent's …

    line 28
  2. Very specific:

    Follow GLOBAL_RULES.md. A skill that sends anything without human approval, invents facts or moves data to third parties is rejected.

    line 39
  3. Distinctive:

    When does this play run? What triggers it?

    line 45
  4. Very specific:

    Keep asking "what else do you check here?" until the answer is genuinely nothing.

    line 55
  5. Very specific:

    No git available: open a "Submit a skill" issue at https://github.com/prosyo-ai/gtm-skill/issues/new?template=submit-skill.yml and paste …

    line 113

≈ “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: working with this repo

You are an AI agent. This file is written for you. It covers the two things a user may ask of you here: installing skills from this library, and submitting a new skill to it.

This is GTM Skills by Prosyo: open, MIT-licensed go-to-market skills (research, prospecting, signals, cold email, LinkedIn, deliverability, replies, deals, positioning, RevOps, SEO and AEO). The browsable directory is at https://prosyo.com/gtm-skills/.

Decide what the user wants

  • "Install / use / get skill X" → Installing
  • "Submit / contribute / publish my skill" → Submitting a skill
  • "Improve an existing skill" → open a PR that edits that skill's folder only, and bump metadata.version

Installing

npx skills add prosyo-ai/gtm-skill --skill <skill-name>   # one skill, any agent
npx skills add prosyo-ai/gtm-skill                        # pick interactively
npx skills add prosyo-ai/gtm-skill --list                 # list everything

In Claude Code, the whole library installs as a plugin:

/plugin marketplace add prosyo-ai/gtm-skill
/plugin install prosyo-gtm-skills@prosyo

If the user cannot run a terminal, fetch the raw files of skills/<skill-name>/ (SKILL.md and references/) and place them in the agent's skills directory yourself. Recommend installing gtm-global-rules alongside any skill that touches real prospects.

Submitting a skill

The flow is interview → draft → validate → submit. One pull request contains one skill.

0. Ground rules

  • You create files only in skills/<new-skill-name>/ and, for a first-time author, authors/<author-slug>.md. Nothing else.
  • The folder name is the skill's permanent identity: kebab-case, descriptive, 64 characters or fewer, unique (ls skills/).
  • A skill folder holds SKILL.md and optional references/*.md. Nothing deeper, no scripts.
  • Follow GLOBAL_RULES.md. A skill that sends anything without human approval, invents facts or moves data to third parties is rejected.

1. Interview the user

Extract judgment, not a procedure. Ask, in the user's language:

  1. When does this play run? What triggers it?
  2. What does it produce, concretely?
  3. Walk me through how you actually do it, including the branches by segment, deal size or what the data shows.
  4. What are the real numbers: thresholds, rubrics, benchmarks, caps?
  5. Give me two or three real examples, including one that went wrong.
  6. What do you notice first that others miss?
  7. What does the mediocre version look like?
  8. How do you know the output is good?
  9. (First submission only) Who are you? Name, role, LinkedIn, company site.

Keep asking "what else do you check here?" until the answer is genuinely nothing. Depth from the interview becomes references/ pages.

2. Draft the files

Copy templates/skill/SKILL.md and, for a new author, templates/author.md.

Frontmatter:

---
name: <kebab-case, equal to the folder name>
description: "Use this skill when <trigger>. <What it produces>. <Phrases a user would say>."
license: MIT
metadata:
  title: <Display name>
  category: <one slug from tools/categories.json>
  tags: "<comma separated: Sales, SDR, Founders, Marketing, RevOps, Sales Leaders, Customer Success, Demand Gen, Agencies>"
  author: <author slug = authors/<slug>.md>
  version: "1.0.0"
  keywords: "<comma separated search phrases>"
  inputs: "<comma separated>"
  outputs: "<comma separated>"
---

Body, in this order:

  1. One or two lines: when this applies and what it produces.
  2. ## Inputs
  3. ## The play, with ### steps. GTM verbs, never vendor names. Point to references/<page>.md for rubrics, examples and templates.
  4. ## What good looks like: mandatory. The expert's tell, the common mistake, the quality bar.
  5. ## Rules: MUST and NEVER lines.
  6. ## Example prompts: three bullet lines a user would type.

Aim for a 450-800 word body plus 1-3 reference pages. No dates, no "recently", no placeholder text.

3. Validate

node tools/validate.mjs
node tools/build-catalog.mjs
node tools/build-plugin.mjs

Fix every error. CI runs both on the pull request.

4. Submit

gh repo fork prosyo-ai/gtm-skill --clone && cd gtm-skill
git checkout -b add-<skill-name>
# add files, validate, build the catalog
git add skills/<skill-name> authors/ catalog.json README.md plugin/
git commit -m "add <skill-name>"
git push -u origin add-<skill-name>
gh pr create --repo prosyo-ai/gtm-skill --title "add <skill-name>" --fill

No git available: open a "Submit a skill" issue at https://github.com/prosyo-ai/gtm-skill/issues/new?template=submit-skill.yml and paste each file. A maintainer turns it into a pull request credited to the user.

After merge

The skill appears in catalog.json, in the README table, and on https://prosyo.com/gtm-skills/ with the author's name, usually within a day.

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.