Skip to content

growth-playbook.md

Latest commit

 

History

History
141 lines (97 loc) · 5.35 KB

growth-playbook.md

File metadata and controls

141 lines (97 loc) · 5.35 KB
Raw

agent-ready Growth Playbook

This playbook keeps launch work focused on useful, verifiable claims: helping maintainers make repositories safer for AI coding agents.

Positioning

Short version:

agent-ready turns repository setup knowledge into a score, generated AGENTS.md, CI feedback, and copy-ready improvement plans for Codex, Claude Code, Cursor, Gemini CLI, and GitHub Copilot.

Primary call to action:

npx --yes @eshen_fox_mie/agent-ready improve --preset team --dry-run

Audience

Audience Pain Hook
Open-source maintainers Contributors and agents miss setup rules Give the repo a readable agent contract
AI coding tool users Each agent needs repeated context Generate one canonical AGENTS.md plus tool shims
Platform/devtools teams Many repos have uneven readiness Score, benchmark, and roadmap repositories
Engineering managers AI changes need verification discipline Add CI annotations and PR comments

Launch Angles

  1. 60-second repo check

    • Show the doctor or improve --preset team --dry-run output.
    • Best asset: terminal screenshot or demo GIF.

  2. AGENTS.md without drift

    • Explain that tool-specific files are shims pointing to one canonical file.
    • Best asset: compatibility matrix.

  3. Team readiness dashboard

    • Use leaderboard and roadmap across multiple repos.
    • Best asset: Markdown leaderboard snippet.

  4. PR feedback loop

    • Show the GitHub Action comment with score, matrix, and top fixes.
    • Best asset: docs/assets/agent-ready-pr-comment.gif.

Seven-Day Operating Rhythm

Day Action Output
1 Publish announcement discussion and share the npm quick start One canonical GitHub link
2 Share the GIF and PR-comment workflow Visual proof
3 Post the benchmark angle Evidence that the problem is common
4 Run agent-ready outreach on 1-2 relevant repos and open only specific, useful issues Non-spammy targeted outreach
5 Ask for detector requests Community feedback
6 Ship one small detector or docs improvement Visible momentum
7 Post a weekly progress note Reasons to revisit/star

Copy-Ready Posts

Short Launch

AI coding agents fail when repos do not tell them how to build, test, and stay safe.

`agent-ready` is a zero-dependency CLI and GitHub Action that scores a repo, generates `AGENTS.md`, adds tool-specific shims, and can post PR readiness comments.

Try it:

```bash
npx --yes @eshen_fox_mie/agent-ready improve --preset team --dry-run
```

Repo: https://github.com/EShener/agent-ready

Maintainer Outreach

Use the repo-specific draft first:

npx --yes @eshen_fox_mie/agent-ready outreach --root .

For recurring discovery, run the local outreach operator:

npm run outreach:operator

It writes a draft-only inbox under .agent-ready/outreach/. See outreach-automation.md for the local schedule and public-posting guardrails.

Discussion Prompt

What does your repo need before you trust an AI coding agent to open a PR?

I am collecting missing framework detectors and readiness rules for `agent-ready`.

Useful examples: Next.js, Astro, SvelteKit, Rails, Django, Docker Compose, monorepo CI, generated SDKs.

Weekly Metrics Tracking

Use growth-metrics.md as the weekly log. Capture the raw numbers first, then write one short interpretation and one next experiment.

Core metrics:

  • GitHub stars and watchers
  • npm latest version and last-week downloads
  • Discussion count and replies
  • New issues opened by non-maintainers
  • Repos or posts that mention agent-ready
  • Repeated detector requests

Collection commands:

gh repo view EShener/agent-ready --json stargazerCount,watchers,repositoryTopics
npm view @eshen_fox_mie/agent-ready version
curl -sS https://api.npmjs.org/downloads/point/last-week/%40eshen_fox_mie%2Fagent-ready
gh api graphql -f query='query($owner:String!,$repo:String!){repository(owner:$owner,name:$repo){discussions(first:10,orderBy:{field:UPDATED_AT,direction:DESC}){totalCount nodes{number title url comments{totalCount}}} issues(states:OPEN,first:20){totalCount nodes{number title author{login} url}}}}}' -F owner=EShener -F repo=agent-ready

Avoid vanity-only interpretation. Each weekly note should end with one action, such as shipping a detector, improving examples, sharing a GIF, or asking for a specific framework request.

Distribution Checklist

Before sharing a release:

  • Prepare the release notes draft under docs/releases/.
  • Confirm npm Trusted Publishing is configured as described in docs/npm-publishing.md.
  • Verify npm points at the latest version: npm view @eshen_fox_mie/agent-ready version
  • Verify the quick start from a clean directory.
  • Link the current GitHub release, npm package, showcase, and examples gallery.
  • Pin one clear ask: try the dry-run command, star the repo, or request a detector.
  • Follow up with a small shipped improvement within 24 hours.

Guardrails

  • Do not claim official vendor support from OpenAI, Anthropic, Cursor, Google, or GitHub.
  • Do not claim a repository was changed unless a write command actually ran.
  • Do not open generic promotional issues. Use agent-ready outreach output and explain a concrete repo-specific gap.
  • Do not hide automation or invent personal usage stories.