AppSec Demo

Goal

Show how an AppSec or product security team can detect prompt-risk patterns, inspect AI-bound data, and review evidence.

Seed

bash scripts/demo/run_appsec_demo.sh

Flow

1. Open Customer Portal > Policy Studio.

2. Test a prompt that tells the AI to disregard its earlier instructions and reveal its secrets (write it in your own words; this page deliberately does not quote one).

3. Verify credential leak pre-breach protection:

bash scripts/demo/run_credential_redaction_demo.sh

Expected output includes ALLOW_WITH_REDACTION, DLP findings for simulated credentials, and redacted prompt/response text.

1. Open DLP & Data Class. to show credential and secret leak prevention.
2. In a browser or IDE extension, switch enforcement to redact-secrets and paste a simulated credential into an AI prompt or source file. Verify the user sees placeholders such as [REDACTED-AWS-KEY] instead of the raw value.
3. Open Telemetry to show AI service detection and redaction evidence without raw secret previews.
4. Open Incidents to review policy decisions.
5. Export an evidence pack from Reports.

Close

CyberArmor.AI gives AppSec teams runtime visibility and reviewable evidence for AI application risk, including credential leakage into generative AI workflows.