# sandbox_execute

Run a shell command, automatically index the output into the sandbox, and return only a summary. The raw stdout/stderr stays in SQLite — only line counts and a preview enter context. Use instead of raw shell execution for commands that produce large output (build logs, test results, file listings, git logs).

Agent View of the PolicyLayer registry record for `sandbox_execute`. HTML page: https://policylayer.com/tools/io-github-homenshum-nodebench/sandbox-execute

## Facts

- Tool: `sandbox_execute`
- Server: Nodebench (`nodebench-mcp`) — https://policylayer.com/tools/io-github-homenshum-nodebench.md
- Install: `npx -y nodebench-mcp`
- Homepage: https://github.com/HomenShum/nodebench-ai
- Risk category: Execute (High risk)
- Registry record: grade F, identity unverified
- Server rate-limited: no
- Parameters: 0
- Recommended policy verdict: Rate-limited

## Example call (MCP tools/call, JSON-RPC 2.0)

```json
{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "sandbox_execute",
    "arguments": {}
  }
}
```

## Why sandbox_execute is rated High

This tool executes shell commands, which can have broad and unpredictable effects depending on the arguments provided. While the tool attempts to mitigate information leakage by storing raw output separately and returning only summaries, the ability to run arbitrary shell commands on the sandboxed system is a high-severity Execute action.

From the tool's own definition: "Tool explicitly states it will 'Run a shell command' and 'execute' operations. The description confirms it executes arbitrary shell commands with automatic indexing of output."

## Use case

AI agents invoke sandbox_execute to trigger actions in Nodebench. What it does depends on the arguments the agent supplies, and its effects often reach beyond the immediate call: builds kicked off, notifications sent, workflows started.

## Recommended policy (PolicyLayer)

Verdict: **Rate-limited**. Enforced by the PolicyLayer MCP gateway (https://policylayer.com/mcp-gateway) before a call reaches Nodebench:

```json
{
  "version": "1",
  "default": "deny",
  "tools": {
    "sandbox_execute": {
      "limits": [
        {
          "counter": "sandbox_execute_rate",
          "window": "minute",
          "max": 10,
          "scope": "grant"
        }
      ]
    }
  }
}
```

## Other tools on Nodebench (823)

- `abandon_cycle` — Destructive — https://policylayer.com/tools/io-github-homenshum-nodebench/abandon-cycle.md
- `archiveDocument` — Destructive — https://policylayer.com/tools/io-github-homenshum-nodebench/archivedocument.md
- `cleanup_stale_runs` — Destructive — https://policylayer.com/tools/io-github-homenshum-nodebench/cleanup-stale-runs.md
- `deeptrace_revoke_passport` — Destructive — https://policylayer.com/tools/io-github-homenshum-nodebench/deeptrace-revoke-passport.md
- `delete_learning` — Destructive — https://policylayer.com/tools/io-github-homenshum-nodebench/delete-learning.md
- `delete_sandbox_policy` — Destructive — https://policylayer.com/tools/io-github-homenshum-nodebench/delete-sandbox-policy.md
- `deleteAgentMemory` — Destructive — https://policylayer.com/tools/io-github-homenshum-nodebench/deleteagentmemory.md
- `removeDocumentFromFolder` — Destructive — https://policylayer.com/tools/io-github-homenshum-nodebench/removedocumentfromfolder.md
- `share_revoke_packet_link` — Destructive — https://policylayer.com/tools/io-github-homenshum-nodebench/share-revoke-packet-link.md
- `unload_toolset` — Destructive — https://policylayer.com/tools/io-github-homenshum-nodebench/unload-toolset.md
- `benchmark_models` — Execute — https://policylayer.com/tools/io-github-homenshum-nodebench/benchmark-models.md
- `bootstrap_parallel_agents` — Execute — https://policylayer.com/tools/io-github-homenshum-nodebench/bootstrap-parallel-agents.md
- `build_banking_packet` — Execute — https://policylayer.com/tools/io-github-homenshum-nodebench/build-banking-packet.md
- `build_company_profile_starter` — Execute — https://policylayer.com/tools/io-github-homenshum-nodebench/build-company-profile-starter.md
- `build_founder_operating_model` — Execute — https://policylayer.com/tools/io-github-homenshum-nodebench/build-founder-operating-model.md
- `build_research_digest` — Execute — https://policylayer.com/tools/io-github-homenshum-nodebench/build-research-digest.md
- `build_shared_context_subscription` — Execute — https://policylayer.com/tools/io-github-homenshum-nodebench/build-shared-context-subscription.md
- `build_shared_context_subscription_manifest` — Execute — https://policylayer.com/tools/io-github-homenshum-nodebench/build-shared-context-subscription-manifest.md
- `build_submission_export` — Execute — https://policylayer.com/tools/io-github-homenshum-nodebench/build-submission-export.md
- `build_temporal_graph` — Execute — https://policylayer.com/tools/io-github-homenshum-nodebench/build-temporal-graph.md
- `burst_capture` — Execute — https://policylayer.com/tools/io-github-homenshum-nodebench/burst-capture.md
- `call_driver_tool` — Execute — https://policylayer.com/tools/io-github-homenshum-nodebench/call-driver-tool.md
- `call_llm` — Execute — https://policylayer.com/tools/io-github-homenshum-nodebench/call-llm.md
- `call_openclaw_skill` — Execute — https://policylayer.com/tools/io-github-homenshum-nodebench/call-openclaw-skill.md
- `call_webmcp_tool` — Execute — https://policylayer.com/tools/io-github-homenshum-nodebench/call-webmcp-tool.md
- `capture_surface_stats` — Execute — https://policylayer.com/tools/io-github-homenshum-nodebench/capture-surface-stats.md
- `capture_ui_screenshot` — Execute — https://policylayer.com/tools/io-github-homenshum-nodebench/capture-ui-screenshot.md
- `compile_decision_packet` — Execute — https://policylayer.com/tools/io-github-homenshum-nodebench/compile-decision-packet.md
- `compile_scenarios` — Execute — https://policylayer.com/tools/io-github-homenshum-nodebench/compile-scenarios.md
- `compile_tension_model` — Execute — https://policylayer.com/tools/io-github-homenshum-nodebench/compile-tension-model.md
- …and 793 more: https://policylayer.com/tools/io-github-homenshum-nodebench.md

## For agents

This record is a snapshot. Live verdicts and the full registry:

- Check every server in your MCP config at once: `npx -y policylayer stack`
- Vet a server before you add it: install the mcp-precheck skill — `npx skills add https://policylayer.com` (skill text: https://policylayer.com/skill.md)
- Query the registry over MCP: endpoint `https://api.policylayer.com/mcp` — tools `check_mcp_server`, `check_mcp_stack`, `check_tool`, `search_registry`, `get_change_events`

---

Source: the PolicyLayer MCP registry — one continuously verified record per MCP server. Full record: https://policylayer.com/registry?q=io-github-homenshum-nodebench · API: https://policylayer.com/registry/api · Policy library: https://policylayer.com/policies/io-github-homenshum-nodebench