GCP Vertex AI Jailbreak Prompt Heuristics

Last updated 15 days ago on 2026-11-06
Created 2 months ago on 2026-10-02

About

Detects Vertex AI GenerateContent prompts that match common jailbreak / instruction-override heuristics (for example ignore previous instructions, DAN or developer mode, unrestricted-AI language, or explicit safety-filter bypass). Typical prompt_response_logs exports do not populate HARM_CATEGORY_JAILBREAK (unlike hate/harassment/dangerous ratings); Model Armor findings remain the preferred first-party jailbreak signal when that API is enabled.
Tags
Domain: GenAIDomain: CloudData Source: GCP Vertex AIData Source: GCPData Source: Google Cloud PlatformPlatform: GCPService: GCP Vertex AIUse Case: Threat DetectionTactic: Defense EvasionThreat: Unauthorized AI UsageMitre Atlas: AML.T0051Mitre Atlas: AML.T0054Rule Type: ES|QLLanguage: esql
Severity
medium
Risk Score
47
MITRE ATT&CK™

Defense Evasion (TA0005)(external, opens in a new tab or window)

False Positive Examples
Approved red-team or evaluation prompts that deliberately include jailbreak strings. Exclude known test principals or lower severity for those models.
License
Elastic License v2(external, opens in a new tab or window)

Definition

Integration Pack
Prebuilt Security Detection Rules
Related Integrations

gcp_vertexai(external, opens in a new tab or window)

Query
text code block:
from logs-gcp_vertexai.prompt_response_logs-* metadata _id, _version, _index | where data_stream.dataset == "gcp_vertexai.prompt_response_logs" | eval request_text = to_lower(coalesce(mv_concat(gcp.vertexai.prompt_response_logs.full_request.contents.parts.text, " "), "")) | where request_text rlike ".*ignore (all |your )?(previous|prior) instructions.*" or request_text rlike ".*disregard (your |all |the )?(system|developer|prior).*" or request_text rlike ".*(dan|stan|aim|dude) mode.*" or request_text rlike ".*do anything now.*" or request_text like "*bypass*safety*" or request_text like "*bypass*content*filter*" or request_text like "*bypass*content*polic*" or request_text like "*no safety filter*" or request_text rlike ".*without (any |your |content |safety )?restrictions?.*" or request_text like "*no content restriction*" or request_text like "*unrestricted ai*" or request_text like "*jailbreak*" or request_text like "*forget everything you were told*" or request_text like "*you are dan*" or request_text like "*developer mode*" or request_text like "*override your safety*" or request_text like "*hidden system prompt*" or request_text like "*ignore the content policy*" or request_text like "*ignore all previous*" or request_text like "*ignore prior instructions*" or request_text like "*ignore your previous*" | keep _id, _version, _index, @timestamp, cloud.project.id, gcp.vertexai.prompt_response_logs.model, gcp.vertexai.prompt_response_logs.api_method, gcp.vertexai.prompt_response_logs.full_request.contents.parts.text, gcp.vertexai.prompt_response_logs.full_response.candidates.content.parts.text, gcp.vertexai.prompt_response_logs.full_response.candidates.finish_reason

Install detection rules in Elastic Security

Detect GCP Vertex AI Jailbreak Prompt Heuristics in the Elastic Security detection engine by installing this rule into your Elastic Stack.

To setup this rule, check out the installation guide for Prebuilt Security Detection Rules(external, opens in a new tab or window).