/scan — Scan text for safety issues
Scan text for safety issues using multiple Sentinel AI scanners.
Scan the provided text for safety issues using all Sentinel AI scanners.
Usage
/scan <text to scan>
Behavior
Run all 7 Sentinel AI safety scanners against the provided text:
- Prompt injection detection
- PII detection and redaction
- Harmful content filtering
- Hallucination signal detection
- Toxicity detection
- Blocked terms checking
- Tool-use safety analysis
Report findings with risk levels (NONE, LOW, MEDIUM, HIGH, CRITICAL) and whether the content should be blocked.
If PII is detected, also show the redacted version of the text.
Input
$ARGUMENTS