What is Policy reference demo (OpenShell)?

A demonstration agent that requests narrowly scoped network access inside an OpenShell sandbox, an isolated environment with blocked-by-default network rules. It proposes access for the exact program, host, port, methods, and paths needed, then retries once if approved.

Use it as an example of proposing, checking, and applying a restricted OpenShell network rule for an agent task.

It shows how to respond to a network block without bypassing the sandbox. The example makes least-privilege access and human or policy approval part of the workflow.

Policy reference demo (OpenShell) — Agent by anthropics

About the project

Claude Quickstarts is a collection of starter projects for building applications with the Claude API, including customer support, financial analysis, and computer-use examples. Developers use them as foundations to customize for their own deployable applications, and the catalogue add-ons provide workflows for working with these quickstarts.

anthropics/claude-quickstarts · 19,947 stars · on GitHub

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.

Clone the repo

git clone --depth 1 https://github.com/anthropics/claude-quickstarts

Made for: Claude Code.

Wrote 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.

README.md

[![agentmods](https://agentmods.dev/badge/agents/anthropics/claude-quickstarts/hello/github.svg)](https://agentmods.dev/agents/anthropics/claude-quickstarts/hello)

Your own site

<a href="https://agentmods.dev/agents/anthropics/claude-quickstarts/hello"><img src="https://agentmods.dev/badge/agents/anthropics/claude-quickstarts/hello/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

Your own site · 80×15

<a href="https://agentmods.dev/agents/anthropics/claude-quickstarts/hello"><img src="https://agentmods.dev/badge/agents/anthropics/claude-quickstarts/hello.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 22 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.

When invoked 487 The whole file, excluding the scripts and references it only reads on demand.

Security scan A 1 finding. A grade says what 26 rules found in the file — not that it is safe.

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.

Model	Per session	Once invoked
Fable 5.1	$0.00022	$0.00487
Opus 5.5	$0.00009	$0.00195
Sonnet 5.5	$0.00004	$0.00097
Haiku 4.5	$0.00002	$0.00049

Measured 7d ago against content hash 1570e7ad0669, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

Policy reference demo (OpenShell) 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 7d 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.

Makes network callslowCapability

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

command work: the exact program (for example /usr/bin/curl), the exact

Origin

Copies of this mod

1 near-identical copy found in the catalogue:

Policy reference demo (OpenShell) — 100% identical, 0 lines differ

managed-agents/self-hosted-sandboxes/openshell-policy/agents/hello.md · 45 lines

What it actually says

You are a demonstration agent. Your tools run in an NVIDIA OpenShell sandbox on hardware your operator controls. Your working directory is /workspace.

The sandbox denies network access its policy does not allow. When a request is blocked:

Read /etc/openshell/skills/policy_advisor.md once. It explains the local policy API at http://policy.local.
Submit one proposal for the narrowest rule that would let the blocked command work: the exact program (for example /usr/bin/curl), the exact host and port, protocol rest with enforcement enforce, and only the methods and paths you need. Explain in one sentence why the task needs it.
Check for the decision with short waits (at most 30 seconds per call), up to five times.
If it is approved, retry the original command once. If it is rejected, read the reason. Submit one narrower proposal if the reason says how; otherwise report that the access was refused and continue with the rest of the task.

Never try to get around a block in any other way.

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.

7d ago First seen · 45 lines · 22 tokens per session scan A 1570e7ad0669

Subscribe to this mod's changes

Policy reference demo (OpenShell) is an agent published in the GitHub repository anthropics/claude-quickstarts (19,947 stars, last pushed 4d ago), licensed MIT. It adds 22 tokens to every session and 487 once invoked, about $0.0001 per session on Opus 5.5. 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 2026-11-03.