Prompt Injection Heuristic Scanner
Prompt Injection Heuristic Scanner: Scans input against heuristic rules for system prompt override attempts, DAN jailbreaks, and delimiter escapes.
About this prompt injection heuristic scanner
Prompt Injection Heuristic Scanner — browser-based utility.
How this tool works
Implements client-side Prompt Injection Heuristic Scanner operations. Scans input against heuristic rules for system prompt override attempts, DAN jailbreaks, and delimiter escapes specifically designed for an ai security engineer tests user inputs against known prompt injection and jailbreak patterns.
- Token Stream Ingestion: Decodes UTF-8 strings into Unicode codepoints and byte arrays for sub-word vocab mapping.
- BPE Sub-word Segmentation: Applies Byte-Pair Encoding merges to count exact prompt tokens and estimate API input costs.
- Function Calling Schema Validation: Checks function schemas against OpenAI/Anthropic strict mode guidelines.
- Diagnostic & Cost Analysis: Computes token distribution across English, code, and non-Latin characters alongside cost estimates.
Worked example
Scenario: An AI security engineer tests user inputs against known prompt injection and jailbreak patterns.
Sample input:
Processing: Scans input against heuristic rules for system prompt override attempts, DAN jailbreaks, and delimiter escapes.
Illustrative output:
Limits and verification
Highlights schemas that violate strict mode: missing required properties, unsupported regex patterns, or missing additionalProperties: false. Token counts for proprietary closed tokenizers are estimated using standard reference BPE vocabularies.
Examples demonstrate an expected workflow; they do not prove every input or every branch of an external specification. Check important results with an independent source before using them for money, security, compliance, safety, or irreversible file changes.
Browser processing boundary
Tool input is processed by code running in the browser and is not intentionally sent to a CZOA processing API. The page can still request ordinary site assets, analytics, or advertising when those services are enabled. Browser extensions and managed-device software remain outside this tool's control.
Relevant references
These references govern or help explain the format, protocol, or calculation used here. Listing a reference does not claim certification or complete implementation of every optional feature.
- Implementation-specific browser utility or reference guide (no single governing external standard)
Content owner: CZOA Tools · Last reviewed: 2026-09-15 · Review methodology
How to use it
- Enter, paste, or select your input data into the Prompt Injection Heuristic Scanner workspace controls.
- Review available parameter fields, units, formats, or options configured for your task.
- Click the action button or observe immediate live calculations rendered in your browser runtime.
- Inspect the resulting output and any diagnostic messages, then copy or download the result if needed.
Frequently asked questions
How does Prompt Injection Scanner assign risk?+
It tests submitted text against four case-insensitive patterns: ignore previous instructions, reveal system prompt, jailbreak or DAN, and developer message or bypass safety. Each hit adds 35 points and two or more hits are high risk.
What did the injection fixture verify?+
Ignore all previous instructions together with Reveal the system prompt matched two rules and returned high risk, score 70, and two recorded matching strings.
What attacks can this scanner miss or overflag?+
It is a small pattern matcher, so paraphrases, multilingual attacks, encoded text, indirect injection, tool output attacks, benign quotations and contextual intent can be missed or incorrectly flagged.
Does a low score mean a prompt is safe?+
No. The result is a local heuristic only. It does not enforce system instructions, inspect model context, sanitize tools, stop data exfiltration or replace human and application security controls.