Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add instructions/maneja81/openorbit/claude-mdgit clone --depth 1 https://github.com/maneja81/OpenOrbitWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/instructions/maneja81/openorbit/claude-md)<a href="https://agentmods.dev/instructions/maneja81/openorbit/claude-md"><img src="https://agentmods.dev/badge/instructions/maneja81/openorbit/claude-md.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.06082 | $0.06082 |
| Opus 5 | $0.03041 | $0.03041 |
| Sonnet 5 | $0.01216 | $0.01216 |
| Haiku 4.5 | $0.00608 | $0.00608 |
Grade C, and why
OpenOrbit CLAUDE.md scanned grade C with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 4d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Recursive force deletehighDestructive command
rm -rf with a variable or a broad path is one typo away from removing the wrong tree.
rm -rf node_modules/electron/dist node_modules/electron/path.txt && node node_modules/electron/install.js How it starts
The opening of the file, as written. The whole thing — 315 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Project Instructions
Use live code evidence before planning or editing. Make the smallest safe change. Preserve existing behavior unless the user explicitly asks to change it.
Context
OpenOrbit — a desktop app that runs a team of AI agents locally, with a central orchestrator delegating to specialist sub-agents, each with its own tools and access to the user's files, apps, and Google account.
Status: released (v0.1.0) and under active development on develop. ~290 source files across an Electron main process, a preload bridge, and a React renderer.
Four built-in sub-agents ship seeded from electron/main/ai/defaultAgents.json: Cipher (configAgent), Atlas (knowledgeAgent), Explorer (explorerAgent), Chrono (taskAgent). The orchestrator ("Orbit") is singular, not an agents row, and is configured from its own prompt file plus settings.
Tech Stack
| Layer | Technology |
|---|---|
| Language | TypeScript 5, strict, ESM ("type": "module") |
| Shell | Electron 43 |
| Build | electron-vite 5 (main / preload / renderer) + Vite 7 |
| Renderer | React 19 + framer-motion, react-markdown, driver.js |
| Orchestration | @openai/agents SDK + openai client |
| Persistence | better-sqlite3 (WAL, foreign keys on) |
| Validation | zod 4 |
| Test | vitest 4 + jsdom + @testing-library/react |
| Lint | eslint 9 (flat config) + typescript-eslint |
| Packaging | electron-builder (no config file yet — defaults only) |
Project Structure
/
index.html — renderer HTML entry (Vite root)
electron.vite.config.ts — main / preload / renderer builds; injects release defines
vite.config.ts — renderer-only web build (npm run dev:web)
vitest.config.ts — jsdom, @ alias, release-define stubs, exclusions
tsconfig.json — project refs → tsconfig.node.json + tsconfig.web.json
eslint.config.mjs
electron/
main/
index.ts — bootstrap: CSP, userData migration, IPC registration, window
appDirs.ts — app-owned userData dirs + legacy-root migration
csp.ts, devLog.ts
ai/ — agents.ts, provider.ts, mcp.ts, webSearchDaemon.ts,
approvalPolicy.ts, httpTools.ts, documentExtract.ts,
skillDistill.ts, userInfoStore.ts, xlsWorker.ts
prompts/*.md — orchestrator + built-in sub-agent system prompts
tools/ — per-agent tool sets (task, knowledge, explorer, folder, …)
defaultAgents.json — seed rows for the four built-in agents
db/ — index.ts (bootstrap), migrations.ts, one store per table
ipc/ — one module per surface, each exporting register*Handlers()
connectors/ — Google OAuth: account, Gmail, Calendar, Drive, Contacts
security/ — secretStorage.ts (AES-256-GCM), externalUrl.ts
net/urlSafety.ts — SSRF/private-host guards for user-authored HTTP tools
tasks/scheduler.ts — recurring + one-shot task runner
github/skillSearch.ts
preload/index.ts — contextBridge surface exposed to the renderer
src/ — React renderer
main.tsx — root render (ErrorBoundary → AgentsApp)
components/ — atoms / molecules / organisms
hooks/ — one hook per feature surface
lib/ — pure helpers, each with a colocated .test.ts
vendor/tabler-icons/ — bundled icon webfont
globals.css
scripts/releaseInfo.ts — build-time release metadata (no runtime network call)
public/ — audio, video, favicon
dist-electron/ — build output (untracked)
wiki/ — GitHub wiki, separate repo cloned locally (gitignored)
0-cowork/ — agent working state (gitignored)
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 4d ago First seen · 315 lines · 6,082 tokens per session scan C fda0699a3aeb
OpenOrbit CLAUDE.md is an instructions file published in the GitHub repository maneja81/OpenOrbit (10 stars, last pushed 12d ago), licensed MIT. It adds 6,082 tokens to every session, about $0.0304 per session on Opus 5. A static security scan graded it C with 1 finding (recursive force delete). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other instructions, from other repositories
OGAD AGENTS.md
AGENTS.md instructions for off-grid-ai/OGAD, covering off grid ai desktop — agent guide, design — desktop-first, off grid brand, use the screen real estate — desktop density rules, what this app is and stack.
OGAD CLAUDE.md
Claude Code instructions for off-grid-ai/OGAD, covering off grid ai desktop — agent guide, design — desktop-first, off grid brand, use the screen real estate — desktop density rules, what this app is and stack.
Team-X CLAUDE.md
Instructions for Git-Rocky-Stack/Team-X, covering claude.md, inheritance from parent repositories, design system, key contacts and skill routing.
Mnemosyne-Neural-OS AGENTS.md
AGENTS.md instructions for Mnemosyne-OS/Mnemosyne-Neural-OS, covering agents.md — orientation for ai agents, what this is, in two lines, read in this order, what you can and cannot see (open vs sealed) and what you can actually do from here.
madcop AGENTS.md
AGENTS.md instructions for linmy666/madcop, covering agents.md — madcop agent 协作规约, ⚡ 1. 黄金法则: 优先调用 madcop/tools/ 中的现成工具, ⚡ 2. 消息优先级高于协议 (message priority over protocol), ⚡ 3. 长效编程任务规范 (long-horizon programming) and 3.1 状态持久化.
rizzo-pii CLAUDE.md
Instructions for Rizzo-AI-Academy/rizzo-pii, covering claude.md, cos'è questo progetto, ambiente — vincoli critici e non ovvi, mappa della repo and tassonomia: 22 tag (dettaglio in tassonomiatag.md).