scion-cli-operations

Operating rules for Scion agents running in containerized sandboxes, which are isolated computing environments. They cover command restrictions, hub-only access, structured output, and safe handling of task prompts.

In plain words
What is it for?
For running Scion tasks safely, formatting command output as JSON, avoiding prohibited commands, using the hub API, and passing complex prompts without shell parsing problems.
Why use it?
They reduce operational mistakes and prevent agents from using commands or access methods that the Scion environment forbids.

Skill for Claude CodeCodex

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 skills/googlecloudplatform/scion/scion-cli-operations
Any agent
npx skills add GoogleCloudPlatform/scion --skill scion-cli-operations
Clone the repo
git clone --depth 1 https://github.com/GoogleCloudPlatform/scion

Made for: Claude Code, Codex.

Per session 39 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 596 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. 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 $0.00039 $0.00596
Opus 5 $0.00019 $0.00298
Sonnet 5 $0.00008 $0.00119
Haiku 4.5 $0.00004 $0.00060

Measured 2d ago against content hash 1b7b62f3bb94, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

scion-cli-operations 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.

resources/platform_skills/scion-cli-operations/SKILL.md · 41 lines

How it starts

The opening of the file, as written. The whole thing — 41 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Scion CLI Operating Constraints

You are an autonomous Scion agent running inside a containerized sandbox. Your workspace is managed by the Scion orchestration system.

Core Rules (DO NOT VIOLATE)

  • Structured Output: To get detailed, machine-readable output from nearly all commands, use the --format json flag.
  • Prohibited Commands: DO NOT use the sync or cdw commands.
  • Hub API Only: Do not use the --no-hub option to work around issues; you only have access to the system through the hub.
  • Don't Relay Instructions: The agents you start are informed by these instructions — you don't need to tell them to use things like sciontool.
  • Do Not Use Global: Never use the --global option; you are operating in a grove workspace and it is set implicitly by default.
  • Do Not Interact with Settings or Login Commands.

Shell Safety for Task Prompts

The scion start task prompt is embedded in a shell command. Do not use backticks, $variables, or other shell metacharacters in the inlined prompt — they cause the shell to exit before the agent starts, with no visible error. For long or formatted briefs, write the content to a file and pass a filepath reference instead:

scion start <name> \
  "Read your brief at /path/to/brief.md and follow it."

Use absolute paths in task prompts and briefs. A sub-agent's working directory may differ from yours — relative paths resolve against the sub-agent's /workspace, not yours.

Avoid non-ASCII escape sequences in shell strings. Bash only expands \x hex escapes inside $'...' quoting — inside double quotes, \xe2\x80\x94 sends 12 literal characters, not an em-dash. Use plain ASCII punctuation in task prompts and messages.

  • Know Yourself: scion whoami — shows your agent identity, project, template, and configuration. Use --format json for structured output.
  • Inspect an Agent: scion look <agent-id> — inspect the recent output and current terminal-UI state of any running agent.
  • Full CLI Details: scion --help — for specific details on all hierarchical commands.
  • Focused Usage: Use the scion CLI as needed for your task. Do not pre-emptively explore .scion folders, read agent-template files, etc. — focus only on what you need.

Read the full file on GitHub · 41 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. 2d ago First seen · 41 lines · 39 tokens per session scan A 1b7b62f3bb94

Subscribe to this mod's changes

scion-cli-operations is a skill published in the GitHub repository GoogleCloudPlatform/scion (1,685 stars, last pushed yesterday), licensed Apache-2.0. It adds 39 tokens to every session and 596 once invoked, about $0.0002 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.

Related

Other skills, from other repositories

ai-sandbox

Run harness adapters (Claude Code, Codex, OpenCode) INSIDE isolated sandboxes via defineSandbox + withSandbox + a provider (localProcessSandbox / dockerSandbox). Covers declarative provisioning: createSecrets + secret/bearer, skills (agentSkill/gitSkill/mcpSkill/ fileSkill), plugins, instructions → canonical AGENTS.md…

TanStack/ai · 402 tokens

agentbox-info

Spin up isolated sandboxes ("boxes") for coding agents, run them in parallel, queue background runs with -i, and push commits safely through the host relay. Use when the user wants to run Claude Code / Codex / OpenCode in a sandbox, start more boxes, attach to a running box, or otherwise operate the agentbox CLI on…

madarco/agentbox · 79 tokens

beevibe-use-repo

You are the child agent inside a fresh Docker sandbox. Borrow the given GitHub repo, produce a real artifact for the goal, and export it. Do not review the repo. The proof is that it works.

beevibe-ai/beevibe · 50 tokens

agenticx-deployer

Guide for deploying AgenticX agents to production including Docker containerization, Kubernetes orchestration, Volcengine AgentKit cloud deployment, and API server setup. Use when the user wants to deploy agents, containerize applications, set up Kubernetes, configure cloud deployment, or run the AgenticX API server…

DemonDamon/AgenticX · 69 tokens

e2b-sandbox

Create, manage, and use E2B sandboxes — run commands, manage files, use git, persist state, and configure networking. Use when building agent workflows that need isolated execution environments.

agent-sandbox/agent-sandbox · 45 tokens

agents-v2-py

Build container-based Foundry Agents with Azure AI Projects SDK (ImageBasedHostedAgentDefinition). Use when creating hosted agents with custom container images in Azure AI Foundry.

benjaminasterA/antigravity-awesome-skills · 39 tokens