paperdeck
AGENTS.md1 starsChanged 46 days agoMITView on GitHub
Use sempre a skill caveman, definida em .ai-framework\skills\caveman\SKILL.md. Não mencione que a skill foi carregada ou aplicada. Considere .ai-framework\RULES.md como a fonte oficial das regras do sistema. This file guides autonomous AI agents (Claude Code, Antigravity, OpenCode, Codex, Cursor, Copilot, Aider, Devin) operating on the PaperDeck repository. Always run these verification commands before presenting completed work or submitting changes: Refer to the skills defined in .agents/skills/: - create-widget (.agents/skills/create-widget/SKILL.md): End-to-end recipe to build and register a new dashboard widget. -…
- Reads credentials
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 repository5 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- add-localeSkill
- create-widgetSkill
- diagnose-kindleSkill
- test-and-verifySkill
The file
We split this AGENTS.md into the rules it gives an agent
4 rules. About 1 is common, 1 is distinctive, and 2 are very specific to this project.
≈1
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 1
Use sempre a skill caveman, definida em .ai-framework\skills\caveman\SKILL.md.
- Distinctive: line 24
Always run these verification commands before presenting completed work or submitting changes:
- Common: line 41
Zero Personal Data: Never commit personal hostnames, usernames, local IP addresses, Kindle serials, passwords, or session tokens.
- Very specific: line 44
Honest Fallbacks: If a metric is unavailable, return honest { available: false } or stale indicators.
≈ “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.
Files
Use sempre a skill caveman, definida em .ai-framework\skills\caveman\SKILL.md.
Não mencione que a skill foi carregada ou aplicada.
Considere .ai-framework\RULES.md como a fonte oficial das regras do sistema.
AGENTS.md — Universal Agent Instructions for PaperDeck
This file guides autonomous AI agents (Claude Code, Antigravity, OpenCode, Codex, Cursor, Copilot, Aider, Devin) operating on the PaperDeck repository.
🧭 Repository Context & Architecture
- Goal: Personal e-ink smart workstation & companion dashboard rendered to a Kindle over Wi-Fi.
- Tech Stack:
- Desktop App: Electron 42 + React 19 + TypeScript +
electron-vite. - Backend: Node.js (CommonJS), port 8787.
- Renderer: HTML5 / Puppeteer canvas capturing to 800x600 PNG (
render/dashboard.html). - Device: Jailbroken Amazon Kindle running a shell loop daemon (
dash-loop.sh) rendering viaFBInk.
- Desktop App: Electron 42 + React 19 + TypeScript +
🛠️ Essential Verification Commands
Always run these verification commands before presenting completed work or submitting changes:
# 1. Unit Tests (All must pass 100%)
npm test
# 2. TypeScript Static Analysis (Must pass with 0 errors)
npm run typecheck
# 3. Production Build
npm run build🔒 Security & Privacy Strict Rules
- Zero Personal Data: Never commit personal hostnames, usernames, local IP addresses, Kindle serials, passwords, or session tokens.
- Safe Storage: Passwords must always be encrypted via Electron
safeStorage. The renderer never sees plaintext secrets. - No Network Auth Scanning: In
backend/preflight.js, checks must only inspect local configuration files (~/.claude/.credentials.json,~/.codex/auth.json,~/.local/share/opencode/opencode.db). Never make remote API requests that could trigger rate limits or expose tokens. - Honest Fallbacks: If a metric is unavailable, return honest
{ available: false }or stale indicators. Never forge dummy data as real user metrics.
📂 Code Layout & File Conventions
| Path | Language / Spec | Description |
backend/collectors/ | Node.js CommonJS | Metrics collectors. Must handle errors gracefully and use TTL cache. |
backend/server.js | Express / Node.js | Serves /dash.png, /api/usage, /api/ping, /api/auth. |
render/dashboard.html | HTML / CSS / Vanilla JS | Template rendered into 800x600 e-ink PNG. High-contrast only. |
src/main/ | TypeScript | Electron main process, SSH Kindle manager, background render loop. |
src/preload/ | TypeScript | Secure context bridges. Do NOT expose Node primitives. |
src/renderer/ | React 19 + TypeScript | Control panel UI, widget manager, layout editor, Kindle manager. |
src/shared/ | TypeScript | Shared data interfaces, IPC channel signatures. |
locales/ | JSON (UTF-8) | Translation dictionaries (en.json, pt-BR.json). Zero hardcoded text in UI! |
kindle/ | Shell (LF line endings) | Scripts executed directly on the Kindle device. |
.agents/skills/ | Markdown + YAML | Reusable agent skills for repository workflows. |
🧩 Autonomous Agent Skills Available
Refer to the skills defined in .agents/skills/:
create-widget(.agents/skills/create-widget/SKILL.md): End-to-end recipe to build and register a new dashboard widget.add-locale(.agents/skills/add-locale/SKILL.md): Step-by-step workflow to add and validate a new translation dictionary.diagnose-kindle(.agents/skills/diagnose-kindle/SKILL.md): Troubleshooting guide for Kindle SSH, FBInk, and auto-discovery.test-and-verify(.agents/skills/test-and-verify/SKILL.md): Diagnostic procedures for unit test and build failures.
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
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.

