agentos CLAUDE.md

A set of repository instructions for agentOS, a runtime and software platform from Rivet. It defines naming, versioning, package, protocol, security, and ownership rules for work in that codebase.

In plain words
What is it for?
Developing agentOS runtime code, language execution, virtual file systems, registries, client APIs, documentation, and publishing workflows.
Why use it?
It gives coding agents the project’s boundaries and conventions so changes do not introduce conflicting package names, unsupported compatibility layers, or mismatched protocol versions.

Instructions file

Install

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.

agentmods
npx agentmods add instructions/rivet-dev/agentos/claude-md
Clone the repo
git clone --depth 1 https://github.com/rivet-dev/agentos
Per session 4,008 This file is loaded in full into every session.
When invoked 4,008 The same file — it is already loaded in full.
Security scan A 1 finding. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.04008 $0.04008
Opus 5 $0.02004 $0.02004
Sonnet 5 $0.00802 $0.00802
Haiku 4.5 $0.00401 $0.00401

Measured yesterday against content hash 7ec9778fa81e, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

agentos CLAUDE.md scanned grade A 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 yesterday.

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.

Makes network callslowCapability

Not a fault in itself. Listed so you know the mod talks to something, and to what.

gawk, real curl/sqlite/duckdb/vim, …) compiled to `wasm32-wasip1` against a
CLAUDE.md · 302 lines

How it starts

The opening of the file, as written. The whole thing — 302 lines — stays where its author put it; the contents beside it link to each section on GitHub.

agentOS

Always spell the product name agentOS, never AgentOS; do not alter type identifiers such as AgentOSActorConfig.

agentOS owns the runtime, kernel, VFS, language execution, registry packages, ACP/session layer, agentOS client APIs, docs, and publish machinery. agentOS Exec is the JavaScript, TypeScript, and Python execution surface of agentOS.

Boundaries

  • Keep agentOS product versions pinned at 0.0.1 in committed files. Release workflows apply real versions transiently with scripts/publish; never commit release-version rewrites.
  • agentOS-owned npm packages must use the @rivet-dev/agentos-* namespace. Registry software packages must use @agentos-software/*. Never introduce packages under @agentos/*.
  • Call guest environments VMs, not sandboxes, except when referring to a package or public API that already uses the word.
  • The protocol has no backward compatibility guarantee. Client, sidecar, and protocol crates ship in same-version lockstep; update both sides together.
  • Generic runtime and language-execution work belongs here. Do not add a compatibility mirror or a second package namespace for AgentOS language execution.
  • Keep root package.json scripts limited to Turbo orchestration; repo-specific commands belong in justfile recipes or scoped package scripts.
  • agentOS targets native Linux/container execution. Browser support is not needed or supported here: browser sources may remain as dormant reference code, but their entrypoints must stay disabled and they must not enter default builds, CI, publication, or behavioral-parity requirements without a separately approved design.

Security Model

Trust model:

  • Client: trusted, except for code/payloads it submits for execution.
  • Sidecar/runtime: trusted enforcement point. It owns the kernel, VFS, mounts/plugins, socket table, permissions, and resource policy.
  • Executor: untrusted V8 isolate or WASM guest. Assume guest JS/Python/WASM and third-party packages are hostile.

Read the full file on GitHub · 302 lines

Changes

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.

  1. yesterday First seen · 302 lines · 4,008 tokens per session scan A 7ec9778fa81e

Subscribe to this mod's changes

agentos CLAUDE.md is an instructions file published in the GitHub repository rivet-dev/agentos (4,435 stars, last pushed 5d ago), licensed Apache-2.0. It adds 4,008 tokens to every session, about $0.0200 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.