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/pixel-process-ug/agent-office/claude-mdgit clone --depth 1 https://github.com/Pixel-Process-UG/agent-officeWhat 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 | $0.01974 | $0.01974 |
| Opus 5 | $0.00987 | $0.00987 |
| Sonnet 5 | $0.00395 | $0.00395 |
| Haiku 4.5 | $0.00197 | $0.00197 |
Grade A, and why
agent-office CLAUDE.md scanned grade A with 0 findings 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 3d 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.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 140 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md
This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
What This Is
Agent Office is a pixel-art virtual office for AI agent teams. Globally installable CLI tool (npm i -g agent-office && office). Agents have live presence states, sit in themed rooms, support multi-LLM providers (Claude Code, OpenAI, Ollama), and can be managed through a retro-styled UI, Telegram bot, or CLI.
Commands
# CLI (global install)
office # Start server + open browser in current project dir
office init # Interactive setup wizard
office status # Show running state from PID file
office agents # List agents from DB
office stop # SIGTERM via PID file
office --port 8080 # Override port
office --no-open # Skip browser open (Docker/CI)
# Development
npm run dev # Start Vite dev server (http://localhost:4173)
npm run build # Build server + frontend (tsup + tsc + vite)
npm run build:server # Build server modules only (tsup → dist-server/)
npm test # Run all tests (vitest run)
npm run test:watch # Tests in watch mode
npm run typecheck # Type check frontend + server
npm run serve:build # Build + start production server
npm run setup:force # Re-run interactive setup wizard
npm run setup:auto # Non-interactive setup (CI-friendly)
npm run db:generate # Generate Drizzle migrations
npm run db:push # Push schema to database
Run a single test file: npx vitest run src/__tests__/app.test.tsx
Architecture
Frontend: React 18 + TypeScript 5.6 + Vite 5. Single-page app with pixel-art map, animated sprites, room navigation, agent CRUD forms, task assignment, and activity feed.
CLI: bin/office.mjs - globally installable entry point. Resolves process.cwd() as project root, creates .agent-office/ for state, auto-runs init on first use, manages PID file.
Backend: Node.js server (server.mjs for production, Vite plugin in vite.config.ts for dev). Shared API layer in src/server/. Default storage: SQLite via Drizzle ORM (zero-config). Optional: PostgreSQL via DATABASE_URL env var. API key auth via OFFICE_API_KEY. Rate limiting (100 writes/min, 300 reads/min). CSP headers.
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.
- 3d ago First seen · 140 lines · 1,974 tokens per session scan A 0533a00caefe
agent-office CLAUDE.md is an instructions file published in the GitHub repository Pixel-Process-UG/agent-office (39 stars, last pushed 5mo ago), licensed MIT. It adds 1,974 tokens to every session, about $0.0099 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other instructions, from other repositories
pixtuoid copilot-instructions.md
Instructions for IvanWng97/pixtuoid, covering github copilot instructions — pixtuoid, architecture invariants (never break these), conventions and build & test.
pixtuoid rust.instructions.md
Rust coding standards for the pixtuoid workspace.
pixtuoid CLAUDE.md
Instructions for IvanWng97/pixtuoid, covering claude.md, what this is, layout, build & test and workflow.
p3a CLAUDE.md
Claude Code instructions for fabkury/p3a, covering claude.md, project overview, build management, build commands and set pythonutf8 to avoid unicode encoding errors on windows.
p3a AGENTS.md
AGENTS.md instructions for fabkury/p3a: You are a highly proficient embedded systems developer with expertise in ESP-IDF and ESP32-P4. The project is p3a, a Wi-Fi pixel art player. When using idf.py, activate the ESP-IDF v5.5.4 environment with . C:\Espressif\tools\Microsoft.v5.5.4.PowerShellprofile.ps1, then set…
geek-casino-cardputeradv CLAUDE.md
Instructions for chatelp/geek-casino-cardputeradv, covering silicon casino — casino de poche pour m5stack cardputer adv, rôles, décisions actées, équilibrage — où vit quoi and identité visuelle — design system.