Cells / Examples

Private code risk

Classify a code snippet while returning only a bounded risk verdict.

Examplev1.1.0Engineering
TRY THIS CELL

Ask within the boundary.

Example

Use the declared input, choose permitted fields, and inspect the recorded answer.

Request
Return fields & access
Requested return fields

Access expires

Run a request to inspect its result.
Execution details

The visible rules process text in this browser. Policy, lifetime, access and rolling disclosure checks apply. Each explicit execution creates one redacted record. Editing, presets and animation do not execute or spend disclosure allowance.

Task → Permitted answer
Owner-controlled boundary
function total(items) {
  return items.reduce((sum, item) => sum + item.price, 0);
}
↓ Code-text pattern checks

The submitted code is treated as text.

Output boundaryRequested fields only

The source remains inside. The answer crosses only after its checks.

Your result

Run a request to inspect its result.

Your answer and its record will appear together.
private-code-risk / OverviewCapability policy

Use the answer. Keep the source.

Check code for explicit risk patterns.

What this Cell defines

A Cell is a revocable execution boundary. Its Manifest describes the accepted input, allowed result fields, withheld information, request lifetime and cumulative disclosure limit.

How it works

Text patterns flag dynamic execution (eval / Function), shell execution, hardcoded secrets and unsafe HTML. A clean verdict means no listed pattern matched; it is not a security audit. Source code is never executed.

Execution details

This example analyzes supplied text locally in your browser. Its rules are public. It does not connect to an owner’s private source, execute submitted code, or establish a confidential runtime.

Protocol and roles Privacy boundaries