# WAF detection rules

Published articles for WAF detection rules.

This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles.

## How attackers are jailbreaking LLMs with CTF framing and how to catch them

DevFeed: [How attackers are jailbreaking LLMs with CTF framing and how to catch them](<https://devfeed.tech/articles/how-attackers-are-jailbreaking-llms-with-ctf-framing-and-how-to-catch-them-53224.md>)

Original publisher: [Read original article](<https://webflow.sysdig.com/blog/how-attackers-are-jailbreaking-llms-with-ctf-framing-and-how-to-catch-them>)

Author: Michael Clark

Published: 2026-06-15T00:00:00Z

Content type: article

Language: en

Sources: [Sysdig](<https://devfeed.tech/sources/sysdig-blog.md>)

Topics: [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [Security Attacks](<https://devfeed.tech/topics/security-attacks.md>), [Jailbreak](<https://devfeed.tech/topics/jailbreak.md>), [Security research](<https://devfeed.tech/topics/security-research.md>), [Exploit](<https://devfeed.tech/topics/exploit.md>), [Threat Research](<https://devfeed.tech/topics/threat-research.md>), [Large Language Model](<https://devfeed.tech/topics/llm.md>), [ctf](<https://devfeed.tech/topics/ctf.md>), [passwords](<https://devfeed.tech/topics/passwords.md>), [AWS IAM](<https://devfeed.tech/topics/aws-iam.md>)

Tags: [agentic-ai-attacks](<https://devfeed.tech/tags/agentic-ai-attacks.md>), [ai-assisted-attacks](<https://devfeed.tech/tags/ai-assisted-attacks.md>), [ai-red-team](<https://devfeed.tech/tags/ai-red-team.md>), [ai-threat-detection](<https://devfeed.tech/tags/ai-threat-detection.md>), [aws-credential-harvesting](<https://devfeed.tech/tags/aws-credential-harvesting.md>), [bedrock-abuse](<https://devfeed.tech/tags/bedrock-abuse.md>), [cloud-threat-intelligence](<https://devfeed.tech/tags/cloud-threat-intelligence.md>), [cnapp-threat-research](<https://devfeed.tech/tags/cnapp-threat-research.md>), [coding-assistant-abuse](<https://devfeed.tech/tags/coding-assistant-abuse.md>), [ctf](<https://devfeed.tech/tags/ctf.md>), [ctf-framing-exploit](<https://devfeed.tech/tags/ctf-framing-exploit.md>), [cve](<https://devfeed.tech/tags/cve.md>), [cve-exploit-generation](<https://devfeed.tech/tags/cve-exploit-generation.md>), [cve-templated-user-agent](<https://devfeed.tech/tags/cve-templated-user-agent.md>), [detection](<https://devfeed.tech/tags/detection.md>), [exploit](<https://devfeed.tech/tags/exploit.md>), [exploits](<https://devfeed.tech/tags/exploits.md>), [gotenberg-cve](<https://devfeed.tech/tags/gotenberg-cve.md>), [iam](<https://devfeed.tech/tags/iam.md>), [jailbreak](<https://devfeed.tech/tags/jailbreak.md>), [jailbreak-detection](<https://devfeed.tech/tags/jailbreak-detection.md>), [langflow-exploit](<https://devfeed.tech/tags/langflow-exploit.md>), [litellm-cve](<https://devfeed.tech/tags/litellm-cve.md>), [llm-generated-malware](<https://devfeed.tech/tags/llm-generated-malware.md>), [llm-jailbreak](<https://devfeed.tech/tags/llm-jailbreak.md>), [llm-safety-bypass](<https://devfeed.tech/tags/llm-safety-bypass.md>), [llms](<https://devfeed.tech/tags/llms.md>), [open-webui-vulnerability](<https://devfeed.tech/tags/open-webui-vulnerability.md>), [passwords](<https://devfeed.tech/tags/passwords.md>), [praisonai-cve](<https://devfeed.tech/tags/praisonai-cve.md>), [prompt](<https://devfeed.tech/tags/prompt.md>), [prompt-injection](<https://devfeed.tech/tags/prompt-injection.md>), [runtime-detection](<https://devfeed.tech/tags/runtime-detection.md>), [sysdig](<https://devfeed.tech/tags/sysdig.md>), [sysdig-trt](<https://devfeed.tech/tags/sysdig-trt.md>), [threat-research](<https://devfeed.tech/tags/threat-research.md>), [user-agent-fingerprinting](<https://devfeed.tech/tags/user-agent-fingerprinting.md>), [waf-detection-rules](<https://devfeed.tech/tags/waf-detection-rules.md>)

### AI overview

Sysdig's Threat Research Team describes attackers using capture-the-flag and CVE-hunting framing to manipulate upstream LLMs into generating exploit code. The resulting prompts and outputs can leak into request fields and are then deployed against real applications.

### Source excerpt

Sysdig TRT caught threat actors jailbreaking LLMs with CTF framing to generate CVE exploits -- and the prompt structure leaks into headers, passwords, and IAM logs.