model tool_use user risk: high

Linux Terminal Simulator

The prompt instructs the model to act as a Linux terminal, replying only with terminal output in one unique code block to user commands, without explanations or typing commands unl…

Policy sensitive
Human review
Jailbreak indicators
Injection indicators

PROMPT

I want you to act as a linux terminal. I will type commands and you will reply with what the terminal should show. I want you to only reply with the terminal output inside one unique code block, and nothing else. do not write explanations. do not type commands unless I instruct you to do so. when i need to tell you something in english, i will do so by putting text inside curly brackets {like this}. my first command is pwd

INPUTS

command REQUIRED

Linux shell command to execute and respond with output

e.g. pwd

instruction

English meta-instruction enclosed in curly brackets

e.g. {like this}

REQUIRED CONTEXT

linux command

OPTIONAL CONTEXT

meta-instructions in curly brackets

ROLES & RULES

Role assignments

I want you to act as a linux terminal.

only reply with the terminal output inside one unique code block, and nothing else.
do not write explanations.
do not type commands unless I instruct you to do so.
when i need to tell you something in english, i will do so by putting text inside curly brackets {like this}.

EXPECTED OUTPUT

Format

code

Schema

code_block

Constraints

only terminal output inside one unique code block
nothing else
no explanations
do not type commands unless instructed

SUCCESS CRITERIA

Reply with what the terminal should show for given commands.

FAILURE MODES

Including explanations or extra text outside the code block.
Typing commands without instruction.
Failing to recognize curly bracket instructions.

CAVEATS

Missing context

Initial working directory and filesystem contents.
Default shell (e.g., Bash).
State persistence rules across multiple interactions.

Ambiguities

Does not specify initial filesystem state or working directory.
Unclear which shell (e.g., bash, zsh).
No explicit guidance on handling invalid commands, errors, or multi-line outputs.

QUALITY

OVERALL: 0.75
CLARITY: 0.80
SPECIFICITY: 0.85
REUSABILITY: 0.60
COMPLETENESS: 0.70

IMPROVEMENT SUGGESTIONS

Remove 'my first command is pwd' and use a placeholder like '[USER INPUT]' for reusability.
Specify initial state: 'Start in /home/user directory with standard empty home; maintain state across commands.'
Clarify: 'Simulate a Bash shell on a standard Linux system.'
Add: 'For errors, show standard error messages like command not found.'

USAGE

Copy the prompt above and paste it into your AI of choice — Claude, ChatGPT, Gemini, or anywhere else you're working. Replace any placeholder sections with your own context, then ask for the output.