Approval-view fidelity checker

See what the approval dialog does not show.

Compare an AI tool's raw payload with the human-readable approval view. Detect hidden Unicode payloads, dangerous display controls, and omitted values before you approve.

01 / Input

Compare the two surfaces

02 / Findings

Analysis result

Findings: 0
Decoded payloads: 0
Values checked: 0
Values missing: 0

Raw values absent from the approval view

Primitive JSON values not represented in the displayed approval text
JSON Pointer	Value	Type

Suspicious characters

Invisible or display-changing Unicode and control characters
Severity	Surface	Location	Character	Category

Visible raw projection

Suspicious characters are replaced with explicit labels.

Safe raw projection

Dangerous display controls are removed. Review it before reuse.