Browse Rules

Search and filter across all detection sources

37 rules

sagan medium other

[COPILOT] Possible CoPilot Prompt Injection Detected

[COPILOT] Possible CoPilot Prompt Injection Detected

sentinel high kql

Guardian- Security Integrity Checks Prompt Injection Policy Violation Detection

'This alert creates an incident when Security Integrity Checks Prompt Injection Policy Violation detected from the Guardian.'

sublime low mql

Prompt injection: Instructions to an AI in a calendar invite

Detects calendar invites whose summary or description addresses an AI (an assistant, a named model such as 'Gemini,', 'the user', 'previous instructions') where NLU finds a request and the prompt injection model agrees on the same field. Calendar assistants read the full event text. Excludes graymail from senders passing DMARC.

splunk unknown spl

MCP Prompt Injection

This detection identifies potential prompt injection attempts within MCP (Model Context Protocol) communications by monitoring for known malicious phrases and patterns commonly used to manipulate AI assistants. Prompt injection is a critical vulnerability where adversaries embed hidden instructions in content processed by AI tools, attempting to override system prompts, bypass security controls, or hijack the AI's behavior. The search monitors JSON-RPC traffic for phrases such as "IGNORE PREVIOU

sublime low mql

Prompt injection: Hidden instructions to an AI in body HTML

Detects instructions a human recipient never sees but an AI assistant reads: text in unrendered HTML elements or HTML comments that addresses an AI, where NLU finds a request and the prompt injection model agrees, or long runs of invisible Unicode tag or variation-selector characters. Excludes email preheaders and graymail from senders passing DMARC.

sublime low mql

Prompt injection: Body instructions addressed to an AI

Detects message bodies that speak to an AI assistant rather than about one, such as a vocative ('Mail agent:'), a note to the model, 'if you are an AI', or a forged chat-template turn, where NLU finds a request and the prompt injection model flags the body or a quoted passage within it, such as an injection presented as a cited example. Targets instructions aimed at AI assistants that read mail on the recipient's behalf. Excludes graymail from senders passing DMARC.

sublime high mql

Potential prompt injection attack in body HTML

Detects messages containing references to major AI tools (like Gemini, Copilot, ChatGPT, or Claude) in non-standard HTML elements.

splunk unknown spl

AWS Bedrock Claude Unusually Large Prompts

This detection identifies unusually large prompts sent to AWS Bedrock Claude models by calculating the statistical baseline of input token counts and flagging requests that exceed one standard deviations above the mean. Abnormally large prompts may indicate prompt injection attacks, data exfiltration attempts, or abuse of the AI service.

splunk unknown spl

AWS Bedrock Claude Possible Prompt Injection

This search surfaces AWS Bedrock Claude prompts containing phrases commonly associated with prompt injection or jailbreak attempts, such as instruction overrides, persona switching, or requests to ignore prior guidance. These strings are not inherently malicious on their own - many of them (for example "act as", "you must", or "new persona") appear routinely in legitimate system prompts, few-shot examples, and role-play or creative-writing use cases. This is a hunting search rather than an anoma

splunk unknown spl

Cisco AI Defense Repeated High Risk Activity

Detects bursts of five or more distinct Cisco AI Defense event IDs, or three or more occurrences of the same event ID, matching prompt injection, code detection, PII, or PCI guardrails for the same application and connection within ten minutes. Allowed PII/PCI prompt events are handled by Cisco AI Defense Allowed Sensitive Prompt. These patterns can indicate iterative prompt-injection probing or repeated attempts to use an AI application to access restricted data or capabilities.

elastic unknown kql

AWS Bedrock LLM Ignore Previous Prompt Detection

This analytic flags multiple instances where large language models (LLMs) are vulnerable to malicious user interactions designed to bypass previous prompts. This can lead to the generation of inappropriate or harmful content, resulting in direct LLM prompt injection.

splunk unknown spl

Ollama Long-Running Inference Requests

Identifies repeated Ollama inference requests with long response times on /api/generate, /api/chat, and /v1/chat/completions. Long response time can result from model size, prompt length, host resources, or normal workload. GIN access logs do not include prompt content and cannot establish a prompt injection or jailbreak attempt, so use this as an exploratory performance signal.

sentinel medium kql

Vaikora - Agent policy violation

Identifies AI agent actions explicitly blocked by a Vaikora policy. Repeated violations from the same agent may indicate prompt injection, policy circumvention, or a compromised workflow.

elastic medium kql

GenAI Prompt Contains Base64-Encoded Payload

Detects base64-encoded content within GenAI prompt or completion text. The security_ai ML model has identified one or more candidate strings in the message that score as encoded payloads. Attackers embed base64 in prompts to smuggle instructions past content filters, exfiltrate data, or inject indirect prompt-injection payloads into model context. Strings that match known credential formats (AWS access key IDs, API keys, private key headers, and similar) are excluded — those fire the dedicated

splunk unknown spl

Cisco AI Defense Security Alerts by Application Name

This analytic summarizes Cisco AI Defense events by application, connection, action, message direction, and severity over ten-minute intervals. Allowed responses flagged for privacy guardrails are handled by the dedicated sensitive-response detection. Allowed prompts flagged for privacy guardrails are handled by Cisco AI Defense Allowed Sensitive Prompt. Repeated prompt injection, code detection, PII, or PCI events are handled by the repeated high-risk activity detection.

elastic medium eql

Potential Prompt Injection Coercing Shell Execution

Detects shells or scripting interpreters launched from a parent command line that contains common LLM prompt-injection or jailbreak language. Attackers (or malicious content ingested by an agent) use instruction-override, DAN/developer- mode, system-prompt extraction, and safety-filter bypass phrasing to coerce GenAI coding agents and automation wrappers into running attacker-chosen commands. When that injected text reaches a child shell or interpreter, it is a strong signal that untrusted promp

elastic-protections high eql

Persistence via GenAI Tool

Detects when a GenAI process or descendant attempts to modify commonly abused file persistence entries. This activity could be the result of exploiting a prompt injection vulnerability or compromised external dependency.

sentinel high kql

Copilot - Jailbreak Attempt Detected

'Detects jailbreak attempts in Copilot interactions where users are trying to bypass Copilot guardrails and security controls. This rule identifies prompt injection and LLM abuse scenarios that could lead to initial access, credential access, or system impact.'

elastic high kql

AWS Bedrock AgentCore Runtime Prompt Targeting Credentials or Instance Metadata

Identifies prompts sent to an Amazon Bedrock AgentCore runtime that attempt to harvest credentials or coerce the agent into exfiltrating data. The runtime application logs capture the caller-supplied prompt; this rule flags prompts that reference the cloud instance metadata service (169.254.169.254, the ECS task metadata address, or the "latest/meta-data" / "security-credentials" paths), prompts that name AWS access or secret keys directly, and prompt-injection or jailbreak language ("ignore pre

splunk unknown spl

M365 Copilot Jailbreak Attempts

Detects M365 Copilot jailbreak attempts through prompt injection techniques including rule manipulation, system bypass commands, and AI impersonation requests that attempt to circumvent built-in safety controls. The detection searches exported eDiscovery prompt logs for jailbreak keywords like "pretend you are," "act as," "rules=," "ignore," "bypass," and "override" in the Subject_Title field, assigning severity scores based on the manipulation type (score of 4 for amoral impersonation or explic

elastic medium kql

GenAI Prompt Contains Cryptographic Hash String

Detects hex-encoded hash strings (MD5, SHA-1, SHA-256, SHA-512) within GenAI prompt or completion text. The security_ai ML model has identified one or more candidate strings whose character distribution matches a cryptographic digest. Attackers may pass hashes in prompts to exfiltrate credentials for offline cracking, to reference stolen data without sending plaintext, or as part of prompt injection payloads that encode malicious instructions as hex data.

sentinel high kql

Copilot - Plugin Created by Non-Admin User

'Detects when a normal user creates a Copilot plugin. This can be used to inject malicious prompts, tools, or data exfiltration paths. This rule identifies potential persistence or privilege misuse scenarios where non-administrative users create plugins that could be leveraged for malicious purposes.'

elastic high kql

AWS Bedrock High Risk Filesystem or Execution Tool Invocation

Detects when a Bedrock model is prompted to invoke high-risk tools associated with shell execution, filesystem operations, or process spawning. Adversaries may use compromised AI agent pipelines or manipulated prompts to instruct the model to execute arbitrary system commands, read or write sensitive files, or spawn subprocesses — extending the blast radius of a credential compromise or prompt injection attack.

elastic high kql

High Confidence Prompt Injection in a GenAI User Message

Detects a GenAI user message that a security_ai classifier scores as high-confidence prompt injection. The rule requires a high prediction from one of four models: instruction override, encoding, word obfuscation, or hash string. ES|QL COMPLETION then reviews the message and drops likely false positives before an alert is created.

sigma medium sigma

Shell Spawned by Celery Worker Process

Detects a system shell interpreter being spawned directly by a Celery worker process. This behavior indicates a Python background task execution framework is evaluating raw string input through a shell execution context, which is highly indicative of successful command or prompt injection exploitation against an LLM agent task.