XPR Code Sandbox

Execute JavaScript code in a secure sandbox for data processing, computation, and quick expression evaluation without network or filesystem access.

Security Scan

The skill's code and runtime instructions are consistent with a JavaScript sandbox for computation and do not request unrelated credentials or installs, but provenance is unknown and sandbox-escape risks (common with vm-based sandboxes) warrant caution.

Purpose & Capability

Name, description, declared tools (execute_js, eval_expression) and the included src/index.ts align: the implementation uses Node's vm to run JS with explicit globals and blocked require/process/global access. No unrelated env vars, binaries, or installs are requested.

Instruction Scope

SKILL.md limits scope to isolated JS execution and explicitly forbids network/filesystem/imports; the code enforces these restrictions by creating a limited context and undefining require/process/global. However, vm-based sandboxes can be subtle to get correct — the instructions do not permit reading system state or secrets, and the code does not appear to do so, but there remains a non-zero risk of sandbox escape vectors in complex JS environments.

Install Mechanism

No install spec; skill is instruction/code-only and uses Node built-ins. Nothing in the manifest pulls external code or runs remote installers.

Credentials

The skill requires no environment variables, no credentials, and no config paths. The code does not access process.env or other outside credentials. Requested privileges are proportional to the stated purpose.

Persistence & Privilege

always is false; the skill is user-invocable and allows autonomous invocation (platform default). It does not modify other skills or request permanent elevated presence.

Assessment

This skill appears internally consistent for running JavaScript computations in a sandbox and does not ask for credentials or installs. Before installing, consider: (1) provenance — the source/homepage is unknown; prefer skills from known authors. (2) Treat it as untrusted when executing third-party code: test with non-sensitive inputs and small timeouts. (3) Although the code explicitly blocks require/process/global and disables code-generation-from-strings, vm-based sandboxes have historically had escape techniques — avoid sending secrets or sensitive data into the sandbox. (4) If you need stronger guarantees, run similar workloads in a dedicated, isolated execution environment (separate process/container) or seek a sandbox with formal audits.

Code Sandbox

You have sandboxed JavaScript execution tools for computation and data processing:

Full scripts:

• execute_js — run JavaScript code in an isolated V8 sandbox
  • Pass data via the input parameter (JSON) — access it as INPUT in your code
  • Use console.log() to capture intermediate values (returned in logs array)
  • Available globals: JSON, Math, Date, Array, Object, String, Number, RegExp, Map, Set, parseInt, parseFloat, isNaN, isFinite, encodeURIComponent, decodeURIComponent, atob, btoa
  • No network access, no filesystem, no imports — pure computation only
  • Default timeout 5 seconds, max 30 seconds
  • 10MB output limit

Quick expressions:

• eval_expression — evaluate a single JavaScript expression and return the result
  • Use for quick math: "15 * 4500 * 0.01" → 675
  • Date calculations: "new Date().toISOString()"
  • Array operations: "[1,2,3].map(x => x*x)" → [1, 4, 9]

Best practices:

• Use execute_js for multi-step data processing, algorithm testing, code validation
• Use eval_expression for quick math, string ops, date calculations
• Pass large datasets via input parameter rather than embedding in code
• Combine with parse_csv (structured-data skill) for CSV → transform → output workflows
• Combine with store_deliverable to save computed results as job evidence