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/chamuka-inc/vmette/claude-mdgit clone --depth 1 https://github.com/chamuka-inc/vmetteWrote 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/chamuka-inc/vmette/claude-md)<a href="https://agentmods.dev/instructions/chamuka-inc/vmette/claude-md"><img src="https://agentmods.dev/badge/instructions/chamuka-inc/vmette/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 | $0.03645 | $0.03645 |
| Opus 5 | $0.01822 | $0.01822 |
| Sonnet 5 | $0.00729 | $0.00729 |
| Haiku 4.5 | $0.00364 | $0.00364 |
Grade A, and why
vmette 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 5d 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 — 196 lines — stays where its author put it; the contents beside it link to each section on GitHub.
vmette — project instructions
vmette is a headless Linux microVM sandbox for macOS, built on Apple's
Virtualization.framework (VZ) via the objc2-virtualization bindings. It boots
a Linux guest in ~1s and ships as a CLI, a Rust library, a C ABI (cbindgen), a
long-lived daemon, and an MCP server for AI agents. macOS-only by design.
Checks (run before considering work done)
This repo has no make ci target. Run the cargo commands directly:
cargo fmt --all --check # formatting (use `cargo fmt --all` to fix)
cargo clippy --workspace --all-targets # lint — keep at ZERO warnings
cargo test --workspace # unit + integration tests
All three must be clean. make test additionally runs the end-to-end VM smoke
(tests/run.sh, ~20 gates) — it boots real VMs, so it requires a codesigned
binary and macOS.
Workspace layout
Eleven crates (workspace version = 0.1.0, edition 2021, rust-version = 1.80, MIT):
| Crate | Purpose |
|---|---|
crates/vmette |
Core library (lib + cdylib + staticlib). VZ wrapper + Session only; Config, run(), providers, C ABI. Depends on vmette-proto. |
crates/vmette-proto |
Leaf (serde only, no workspace deps). The wire contracts: agent::{Action, ResponseHeader, ScrollDirection} (guest vsock), daemon::{Request, Frame, DesktopRequest, DesktopReply, …} (vmetted socket), geom::Rect, ShareMount. One Rust type per wire shape → drift is a compile error. |
crates/vmette-providers |
Aggregator exposing default_registry() (DirProvider→Squashfs→Tar→Oci, the single load-bearing order). Used by the CLI + daemon so both resolve specs identically. |
crates/vmette-assets |
Shared boot-asset (kernel + initramfs) discovery for the binaries. |
crates/vmette-cli |
vmette CLI binary. Hand-rolled arg parsing (no clap — keeps the binary small). |
crates/vmette-daemon |
vmetted — UNIX-socket dispatcher (tokio). In-process stateless run lane + stateful desktop registry; owns the pixel-settle perception module. |
crates/vmette-daemon-client |
Sync transport for the vmetted desktop socket (connect / lazy-auto-spawn / line framing) — the single owner, shared by the CLI (directly) and the MCP server (via spawn_blocking). |
crates/vmette-mcp |
MCP server (vmette-mcp) exposing vmette to AI agents over stdio. |
crates/vmette-provider-oci |
OCI/Docker image rootfs provider (catch-all for bare refs + oci://); per-registry AuthResolver for private images. |
crates/vmette-provider-squashfs |
Squashfs block-image rootfs provider (squashfs+file://, squashfs+https://, squashfs+http://). Returns BlockImage. |
crates/vmette-provider-tar |
Tarball rootfs provider (tar+https://, tar+http://, tar+file://). |
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.
- 5d ago First seen · 196 lines · 3,645 tokens per session scan A 850ec34fe3bf
vmette CLAUDE.md is an instructions file published in the GitHub repository chamuka-inc/vmette (5 stars, last pushed 2mo ago), licensed MIT. It adds 3,645 tokens to every session, about $0.0182 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 2026-08-31.
Other instructions, from other repositories
microsandbox AGENTS.md
AGENTS.md instructions for superradcompany/microsandbox, covering agents.md, scope, project map, design principles and backward compatibility review.
agent-safehouse AGENTS.md
AGENTS.md instructions for eugene1g/agent-safehouse, covering agent safehouse: llm quick reference, what this project is, fast working rules, core commands and generate policy only (prints temp policy path).
SmolVM AGENTS.md
AGENTS.md instructions for CelestoAI/SmolVM, covering smolvm context, 🚀 project overview, 🧪 development, key commands and release checklist.
sandboxd AGENTS.md
AGENTS.md instructions for tastyeffectco/sandboxd, covering agents.md — operating sandboxd, what this is, prerequisites, install and core api.
meta-qcom AGENTS.md
AGENTS.md instructions for qualcomm-linux/meta-qcom, covering agent guide for meta-qcom, project overview, 1) prerequisites, container runtime smoke test (required order) and 2) recommended environment.
greywall CLAUDE.md
Instructions for GreyhavenHQ/greywall, covering greywall, build & run, test, lint & format and project structure.