agentleFS
vybzaivybzai/spitfire-storybook/AGENTS.md

spitfire-storybook

AGENTS.md0 starsChanged 5 months agoMITView on GitHub

Guidance for any AI coding agent working in this repo — Claude Code, OpenAI Codex, Cursor / Composer 2, Grok CLI, Gemini CLI, Windsurf, Zed, and others. This file is the cross-agent standard (agents.md); Cursor also reads .cursor/rules/spitfire-storybook.mdc. spitfire-storybook is a privacy-first, local-only, open-source tool with two halves of one product: 1. Generation — it writes you a full-length, brutally honest, deeply personalized self-development book from your own connected data, on your own machine, compiled to PDF + EPUB. 2.…

  • 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 32,671 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 repository3 other files it gives its agents.

The file

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

5 rules. About 0 are common, 4 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈4

    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:

    Do not copy raw private content (emails, messages, names, financials, client identities) into any tracked file, and never into the book.

    line 31
  2. Distinctive:

    Use only generic or clearly fictional examples in anything you commit.

    line 33
  3. Distinctive:

    Keep it local-only: don't add cloud calls, telemetry, uploads, or a hosted tier.

    line 98
  4. Very specific:

    Don't read a skill's reference/* files until you reach that step — the SKILL.md files load them on demand.

    line 100
  5. Distinctive:

    Before committing, double-check that no profile, intermediate notes, or real personal content is staged.

    line 102

≈ “Common” is an estimate: at least 1 in 5 of the 41,581 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 — spitfire-storybook

Guidance for any AI coding agent working in this repo — Claude Code, OpenAI Codex, Cursor / Composer 2, Grok CLI, Gemini CLI, Windsurf, Zed, and others. This file is the cross-agent standard (agents.md); Cursor also reads .cursor/rules/spitfire-storybook.mdc.

spitfire-storybook is a privacy-first, local-only, open-source tool with two halves of one product:

  1. Generation — it writes you a full-length, brutally honest, deeply personalized self-development book from your own connected data, on your own machine, compiled to PDF + EPUB.
  2. Narration — it then reads that book to you: any book, EPUB, Markdown, or plain text becomes a chaptered .m4b audiobook (Apple Books-ready) in the "Heart" voice, via a bundled local Kokoro server.

The headline differentiator: it doesn't just write your book, it reads it to you. Everything runs on your hardware. There is no hosted/cloud/SaaS version, and nothing is ever uploaded.


Golden rule (non-negotiable)

Never commit real personal data. This is a public repository.

  • The diagnostic profile produced during generation (profile.md / profile/) is private, git-ignored, and never shipped in the compiled book. Verify this on every build.
  • Do not copy raw private content (emails, messages, names, financials, client identities) into any tracked file, and never into the book.
  • Use only generic or clearly fictional examples in anything you commit.

What's in the repo

skills/
  brutally-honest-book-builder/   # Agent Skill: generation pipeline
  narrate/                        # Agent Skill: narration wrapper
cli/
  narrate.py                      # standalone local audiobook CLI (no agent needed)
  kokoro/                         # bundled local Kokoro TTS server

The two Agent Skills (skills/)

Claude Code and Codex auto-detect Agent Skills from a skills directory. Install by copying:

cp -R skills/* ~/.claude/skills/     # Claude Code
cp -R skills/* ~/.codex/skills/      # OpenAI Codex

Cursor, Grok CLI, Gemini CLI and other agents don't need this copy step — they read this AGENTS.md directly from the repo (and Cursor also reads .cursor/rules/). After install, the agent picks the right skill automatically based on what you ask:

  • brutally-honest-book-builder — generation. A consent-gated, read-only pipeline: Collect → Diagnose → Generate (~30–40 chapters, ~120k+ words) → Build (PDF + EPUB via pandoc). Get explicit, source-by-source consent before reading anything; the private diagnostic profile drives the chapters but never ships. Trigger it by asking, e.g., "build me a personalized book from my data."
  • narrate — narration. Turns text/Markdown/EPUB/a folder of chapters into spoken audio or a chaptered .m4b audiobook locally, in the "Heart" voice (af_heart) via Kokoro. No cloud, no API cost.

The natural end-to-end flow: generate a book with the first skill, then narrate its Markdown chapters with the second.

Standalone CLI (cli/ — no AI agent required)

cli/narrate.py is a dependency-light path (system python3 + ffmpeg only; PIL optional for covers) that does narration with no agent at all. It talks to a local Kokoro server you run from cli/kokoro/ (see cli/kokoro/README.md).

# 1. start the bundled local Kokoro server
python cli/kokoro/server.py            # serves http://localhost:8081

# 2. narrate
python cli/narrate.py say "Hello from Heart."
python cli/narrate.py file notes.md -o notes.mp3
python cli/narrate.py book ./book_dir --title "My Book" --author "A. Writer"

The book command accepts a sorted folder of .md/.txt files or a manifest.json, cleans Markdown into spoken text, synthesizes resumably, and emits a chaptered .m4b with chapter markers and an optional generated cover. Override the endpoint with KOKORO_URL and the voice with KOKORO_VOICE.


Working in this repo

  • Keep it local-only: don't add cloud calls, telemetry, uploads, or a hosted tier. Nothing leaves the user's machine.
  • Don't read a skill's reference/* files until you reach that step — the SKILL.md files load them on demand.
  • Before committing, double-check that no profile, intermediate notes, or real personal content is staged.

Install / docs: https://github.com/vybzai/spitfire-storybook

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.