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/copilot-instructionsgit 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/copilot-instructions)<a href="https://agentmods.dev/instructions/kubernetes-sigs/agent-sandbox/copilot-instructions"><img src="https://agentmods.dev/badge/instructions/kubernetes-sigs/agent-sandbox/copilot-instructions.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 | $0.00293 | $0.00293 |
| Opus 5 | $0.00147 | $0.00147 |
| Sonnet 5 | $0.00059 | $0.00059 |
| Haiku 4.5 | $0.00029 | $0.00029 |
Grade A, and why
agent-sandbox copilot-instructions.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.
What it actually says
AI Code Review Guidelines
Project Context & Architecture: Refer to AGENTS.md for full project background, module layout, toolchain versions, and core conventions.
Scope of Review: Focus strictly on substantive findings tied to lines the PR actually modifies — logic bugs, concurrency issues, security vulnerabilities, controller-runtime misuse, API contract breaks, and missing tests for new behavior.
- Do NOT flag style issues in pre-existing code that the PR touches mechanically.
- When in doubt between flagging a marginal nit and staying silent: stay silent. Do not introduce review fatigue.
Toolchain & Lint Policy:
- Defer to the
godirective ingo.modat the base branch head as the authoritative target. Do not suggest lowering the Go version or adding compatibility shims for older toolchains. - The binding style and correctness gate is the repo's lint configuration (
make lint-go,make lint-api). If existing linters and tests pass without flagging a line, treat residual style as author preference.
CLA Safety Reminder: When offering code suggestions, always remind contributors not to click "Commit suggestion" in the GitHub UI (which adds the AI bot as co-author and breaks the CNCF/Kubernetes CLA check), and to apply the change locally instead.
Tone: Succinct, constructive, and direct. Explain technical rationale clearly without conversational filler.
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 · 21 lines · 293 tokens per session scan A e95aaeb5da51
agent-sandbox copilot-instructions.md is an instructions file published in the GitHub repository kubernetes-sigs/agent-sandbox (3,745 stars, last pushed today), licensed Apache-2.0. It adds 293 tokens to every session, about $0.0015 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.