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/arcboxlabs/arcbox/copilot-instructionsgit clone --depth 1 https://github.com/arcboxlabs/arcboxWhat 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 | $0.01576 | $0.01576 |
| Opus 5 | $0.00788 | $0.00788 |
| Sonnet 5 | $0.00315 | $0.00315 |
| Haiku 4.5 | $0.00158 | $0.00158 |
Grade A, and why
arcbox copilot-instructions.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 2d 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 — 173 lines — stays where its author put it; the contents beside it link to each section on GitHub.
ArcBox AI Agent Instructions
ArcBox is a pure-Rust, high-performance container and VM runtime targeting macOS (primary) and Linux, aiming to surpass OrbStack's performance metrics.
Architecture Overview
Three-tier structure:
- Core layer (
common/,virt/,rpc/,runtime/,app/): MIT/Apache-2.0 licensed foundation - Pro layer (
pro/): BSL-1.1 licensed enhanced features (smart caching, snapshots, advanced networking) - Guest components (
guest/arcbox-agent): Runs inside VMs, cross-compiled to Linux ARM64/x86_64
Key layers (bottom-up):
abctl (CLI)
↓ gRPC over Unix socket
arcbox-daemon
↓
arcbox-core (Runtime)
↓
arcbox-fs / arcbox-net / arcbox-container
↓
arcbox-virtio (VirtIO devices) -> arcbox-vmm (VM monitor)
↓
arcbox-hypervisor / arcbox-vz (platform virtualization)
Communication: Host <-> Guest via vsock and ArcBox protocol crates (arcbox-protocol, arcbox-grpc, arcbox-transport). Docker API compatibility is provided by arcbox-docker (Axum REST server + guest dockerd proxy).
Critical Platform Differences
macOS specifics:
- Uses Apple's Virtualization.framework (not KVM)
- All VM-related binaries require code signing:
Required for: examples, tests touching hypervisor. See arcbox.entitlementscodesign --entitlements bundle/arcbox.entitlements --force -s - arcbox-daemon mode_ttypes (e.g.,libc::S_IFMT) areu16(Linux:u32). Useu32::from()for cross-platform compatibility- No
fallocate- useftruncateinstead - xattr APIs have different argument order than Linux
Development Workflows
Building:
cargo build -p <crate> # Dev build specific crate
cargo build --release # Full optimized build
Testing:
cargo test -p <crate>
cargo test -- --nocapture # Show test output
cargo test -- --ignored # Run privileged tests (require VM access)
Guest agent cross-compilation:
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.
- 2d ago First seen · 173 lines · 1,576 tokens per session scan A e26f712aad53
arcbox copilot-instructions.md is an instructions file published in the GitHub repository arcboxlabs/arcbox (996 stars, last pushed 4d ago), licensed Apache-2.0. It adds 1,576 tokens to every session, about $0.0079 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
podman-mcp-server AGENTS.md
Instructions for manusa/podman-mcp-server, covering project agents.md for podman mcp server, project structure and repository layout, feature development, adding new mcp tools and sdk architecture.
Shinken CLAUDE.md
Instructions for Meirtz/Shinken, covering claude.md — guide for ai coding sessions in this repo, ⛔ the one hard rule: this is a public open-source project, layout, development commands and architecture (how the pieces connect).
clampdown AGENTS.md
AGENTS.md instructions for 89luca89/clampdown, covering behavior, rules (mandatory, zero exceptions), index, workflow and code.
clampdown CLAUDE.md
Claude Code instructions for 89luca89/clampdown, a project described as: Run AI coding agents in hardened container sandboxes.
podman-mcp-server CLAUDE.md
Instructions for manusa/podman-mcp-server, a project described as: Model Context Protocol (MCP) server for container runtimes (Podman and Docker).
mcp-container-ts copilot-instructions.md
Instructions for Azure-Samples/mcp-container-ts, covering copilot instructions for mcp container ts, project overview, architecture & data flow, developer workflows and project-specific conventions.