Agent Sandbox is a Kubernetes extension for managing isolated, stateful workloads that run as single long-lived containers with stable identities and persistent storage. It is intended for AI agent runtimes, reinforcement-learning workloads, and other applications that do not fit ordinary stateless deployments. Its catalogue skills and instructions support operating these sandbox workloads.
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/kubernetes-sigs/agent-sandbox/agents-mdgit clone --depth 1 https://github.com/kubernetes-sigs/agent-sandboxWrote 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/kubernetes-sigs/agent-sandbox/agents-md)<a href="https://agentmods.dev/instructions/kubernetes-sigs/agent-sandbox/agents-md"><img src="https://agentmods.dev/badge/instructions/kubernetes-sigs/agent-sandbox/agents-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 5.1 | $0.04327 | $0.04327 |
| Opus 5 | $0.02164 | $0.02164 |
| Sonnet 5 | $0.00865 | $0.00865 |
| Haiku 4.5 | $0.00433 | $0.00433 |
Grade A, and why
agent-sandbox AGENTS.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 — 143 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md
Guidance for AI coding agents working in this repository. Human contributors should also read CONTRIBUTING.md, docs/development.md, and docs/testing.md, which are the source of truth.
Project summary
agent-sandbox is a Kubernetes controller that provides the Sandbox CRD: a stateful, singleton, pod-backed workload with a stable identity, intended for AI agent runtimes, dev environments, notebooks, and similar use cases. It is a SIG Apps subproject (sigs.k8s.io/agent-sandbox) and follows Kubernetes / controller-runtime conventions.
- API group
agents.x-k8s.io/v1beta1:Sandbox(core). - API group
extensions.agents.x-k8s.io/v1beta1:SandboxClaim,SandboxTemplate,SandboxWarmPool(opt-in extensions). - Go module:
sigs.k8s.io/agent-sandbox(Go 1.26.x, see go.mod).
Repository layout
| Path | What lives there |
|---|---|
| .agents/skills/ | Specialized instructions for you (AI agents). See "Agent Skills" section below. |
| api/v1beta1/ | Core Sandbox types and kubebuilder markers. |
| extensions/api/v1beta1/ | SandboxClaim, SandboxTemplate, SandboxWarmPool types. |
| controllers/ | Core Sandbox reconciler + tests. |
| extensions/controllers/ | Reconcilers for the extension CRDs. |
| cmd/agent-sandbox-controller/ | Controller manager entrypoint. |
| internal/ | Shared internals: lifecycle, metrics, version. Not importable by external consumers. |
| k8s/ | Generated CRDs (k8s/crds/), RBAC, controller manifests. |
| clients/k8s/ | Generated Kubernetes-style clientset, listers, informers (output of dev/tools/client-gen-go.sh). Do not hand-edit. |
| clients/go/ | Hand-written high-level Go SDK that wraps the SandboxClaim lifecycle and exposes Gateway / port-forward / direct connectivity. Editable. |
| clients/python/agentic-sandbox-client/ | Hand-written Python SDK. Directory is named agentic-sandbox-client but the package publishes to PyPI as k8s-agent-sandbox — that's the name to import in docs and examples. |
| examples/ | Runnable sample manifests and demo apps. |
| test/e2e/, test/benchmarks/ | End-to-end and benchmark suites (Go-driven; some scenarios shell out via the Python SDK). |
| dev/tools/ | Repo tooling (lint, generate, deploy-kind, release scripts). Most make targets shell out here. |
| dev/ci/ | Prow presubmit/periodic scripts. |
| docs/ | Development, testing, configuration docs and KEPs (docs/keps/). |
| site/ | Hugo + Docsy source for https://agent-sandbox.sigs.k8s.io. Many pages are thin wrappers that include-file from the repo via mounts in site/hugo.yaml — see "Docs site mounts" below. Native page sources (lifecycle, snapshots, use-cases, runtime-templates, getting started, etc.) live only here. |
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 Changed · +1 lines · +109 tokens per session 2e5cf471e8b7
- 6d ago First seen · 142 lines · 4,218 tokens per session scan A 55364c5c4b8f
agent-sandbox AGENTS.md is an instructions file published in the GitHub repository kubernetes-sigs/agent-sandbox (3,745 stars, last pushed yesterday), licensed Apache-2.0. It adds 4,327 tokens to every session, about $0.0216 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
OpenSandbox AGENTS.md
AGENTS.md instructions for opensandbox-group/OpenSandbox, covering opensandbox agents, repository map, routing, working principles and guardrails.
agentos CLAUDE.md
Claude Code instructions for rivet-dev/agentos, covering agentos, boundaries, security model, sqlite schema ownership and runtime and registry.
OpenSandbox CLAUDE.md
Claude Code instructions for opensandbox-group/OpenSandbox: See AGENTS.md for all rules, routing, and conventions.
ironcurtain CLAUDE.md
Claude Code instructions for provos/ironcurtain, covering claude.md, general workflow, git & worktrees, git workflow and platform considerations.
ironcurtain copilot-instructions.md
Copilot instructions for provos/ironcurtain, covering ironcurtain copilot review instructions, platform support, security invariants, ai sdk v6 api (breaking changes) and code conventions.
ironcurtain security-boundary.instructions.md
Instructions for provos/ironcurtain, covering security boundary review rules, mandatory checks and structural invariant order.