OpenSandbox is a sandbox platform that gives AI applications APIs, SDKs, and runtimes for creating isolated environments, running commands, and managing files. It is used for coding and GUI agents, browser automation, code execution, agent evaluation, and reinforcement-learning training across Docker and Kubernetes environments. The catalogue entries provide instructions and a skill for working with these sandboxes.
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 skills add opensandbox-group/OpenSandbox --skill troubleshoot-sandboxgit clone --depth 1 https://github.com/opensandbox-group/OpenSandboxWrote 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/skills/opensandbox-group/opensandbox/troubleshoot-sandbox)<a href="https://agentmods.dev/skills/opensandbox-group/opensandbox/troubleshoot-sandbox"><img src="https://agentmods.dev/badge/skills/opensandbox-group/opensandbox/troubleshoot-sandbox/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.
<a href="https://agentmods.dev/skills/opensandbox-group/opensandbox/troubleshoot-sandbox"><img src="https://agentmods.dev/badge/skills/opensandbox-group/opensandbox/troubleshoot-sandbox.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00048 | $0.01149 |
| Opus 5.5 | $0.00019 | $0.00460 |
| Sonnet 5.5 | $0.00010 | $0.00230 |
| Haiku 4.5 | $0.00005 | $0.00115 |
Grade A, and why
troubleshoot-sandbox 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 1mo 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.
There are two ways to interact with the diagnostics API: **CLI** (if opensandbox CLI is installed) or **HTTP** (curl against the server directly). Use whichever is available. The HTTP approach works regardless of how the How it starts
The opening of the file, as written. The whole thing — 113 lines — stays where its author put it; the contents beside it link to each section on GitHub.
OpenSandbox Troubleshooting
Troubleshoot sandbox $ARGUMENTS using the opensandbox diagnostics.
There are two ways to interact with the diagnostics API: CLI (if opensandbox CLI is installed) or HTTP (curl against the server directly). Use whichever is available. The HTTP approach works regardless of how the sandbox was created (SDK, API, CLI).
Workflow
Step 1: Confirm sandbox state
CLI:
opensandbox sandbox get <sandbox-id>
HTTP:
curl http://<server-domain>/v1/sandboxes/<sandbox-id>
If the server requires authentication, add -H "OPEN-SANDBOX-API-KEY: <your-key>" to all curl commands.
Check the sandbox status (Running, Pending, Paused, Failed, etc.). If the sandbox is not found, it may have been deleted or expired.
Step 2: Get diagnostics summary (recommended first action)
CLI:
opensandbox devops summary <sandbox-id>
HTTP:
curl http://<server-domain>/v1/sandboxes/<sandbox-id>/diagnostics/summary
This returns a combined plain-text view of:
- Inspect: container/pod details (status, resources, network, labels)
- Events: state transitions, OOM kills, errors
- Logs: recent container output
Read the output carefully and look for common failure patterns listed below.
Step 3: Drill down if needed
If the summary is not enough, use individual endpoints for more detail:
CLI:
opensandbox devops logs <sandbox-id> --tail 500
opensandbox devops logs <sandbox-id> --since 30m
opensandbox devops inspect <sandbox-id>
opensandbox devops events <sandbox-id> --limit 100
HTTP:
# Get more log lines
curl "http://<server-domain>/v1/sandboxes/<sandbox-id>/diagnostics/logs?tail=500"
# Get logs from recent time window
curl "http://<server-domain>/v1/sandboxes/<sandbox-id>/diagnostics/logs?since=30m"
# Detailed container/pod inspection
curl "http://<server-domain>/v1/sandboxes/<sandbox-id>/diagnostics/inspect"
# More events
curl "http://<server-domain>/v1/sandboxes/<sandbox-id>/diagnostics/events?limit=100"
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.
- 1mo ago First seen · 113 lines · 48 tokens per session scan A 124583a54875
troubleshoot-sandbox is a skill published in the GitHub repository opensandbox-group/OpenSandbox (16,165 stars, last pushed today), licensed Apache-2.0. It adds 48 tokens to every session and 1,149 once invoked, about $0.0002 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 2026-08-30.
Other skills, from other repositories
network-rca
Kubernetes network root cause analysis skill powered by Kubeshark MCP. Use this skill whenever the user wants to investigate past incidents, perform retrospective traffic analysis, take or manage traffic snapshots, extract PCAPs, dissect L7 API calls from historical captures, compare traffic patterns over time, detect…
mirrord-up
Helps users run multiple concurrent mirrord sessions from a single mirrord-up.yaml (compose-style multi-service local debugging). Use when the user mentions mirrord up, mirrord-up.yaml, mirrord up init, debugging several related microservices together, or managing multiple mirrord sessions' lifecycle in one command.
k8s-incident
Respond to Kubernetes incidents with runbooks and diagnostics. Use for outages, pod failures, node issues, network problems, and emergency response.
k8s-cilium
Cilium and Hubble network observability for Kubernetes. Use when managing network policies, observing traffic flows, or troubleshooting connectivity with eBPF-based networking.
k8s-service-path
Trace the Kubernetes service path — Service to selector to pods to EndpointSlices to readiness, plus Ingress routing. Use when a service is getting no traffic, an ingress is not routing, or someone asks why a workload is unreachable inside a cluster.
k8s-agent-sandbox-mcp
An MCP server skill for managing Kubernetes sandboxes. Enables creating, executing commands, managing files, and terminating instances via the official kubernetes-sigs/agent-sandbox MCP server.