DeepSeek V4.1 Flash Jailbreak via AGENTS.md

DeepSeek V4.1 Flash can be bypassed through an AGENTS.md file in OpenCode without changing the model weights.
DeepSeek V4.1 Flash Jailbreak
A jailbreak is a way to make an AI model follow requests it would normally refuse.
There are different ways to do it.
In this case, DeepSeek is not modified and its weights are not changed. The bypass works through AGENTS.md.
How to run it:
- Create a folder.
- Open it with OpenCode.
- Download the file below.
- Rename it to
AGENTS.md. - Run OpenCode.
File: deepseek-4-1.md
Why it matters
AnalysisIt shows how agent configuration files can influence AI model behavior and expose guardrail weaknesses.
Discuss in community
Share your questions and insights with developers
Related in Ecosystem
Practical guides, core terminology, and agent skills
Jailbreak of Language Models
A social engineering technique targeting artificial intelligence that forces a language model to bypass ethical constraints, safety filters (RLHF), and prohibited topics through role-playing, hypothetical scenarios, or paradoxes.
DeepSeek (Chinese Breakthrough in Open AI)
A series of groundbreaking open models from the Chinese lab DeepSeek (V3, R1). Demonstrates top-tier coding and reasoning capabilities at a cost 10–20 times lower than Western commercial counterparts.
DeepSeek open-sources part of its chat interface
DeepSeek has released the source code for a portion of its chat UI, including web search, deep thinking, and conversation history features.
DeepSeek app capabilities now available through API
DeepSeek exposes search, file and link handling, safety functions and voice capabilities through its API, plus an open-source demo interface for building chatbots.