DoubleTap

DoubleTap is designed to locate, monitor, and manage large language model (LLM) processes on my Linux x64 system running from a bash shell. It will detect running LLM instance from port default 11434 and name to hunt and kill llama including ollama. Following kill option y/N Shepherd Daemon boot option y/N and uncomplicated fire wall settings with sanity audit to use llama as localhost while preventing outside access as default to keep your local llama for local access. Fencing in your llama for personal use.

doubletap - Local LLM Process Management and Network Tool

Purpose: Emergency termination of resource-hungry LLM processes and associated services
Use Case: When local LLMs consume excessive RAM or maintain unwanted network connections
Kill the llama before it eats all your grass
hint: Always monitor htop for resource usage and netstat for open ports
Maintain proper firewall rules for all AI/ML service ports

default Ollama port 11434 firewall restricted to localhost from uncompicated fire wall

sudo apt install ufw

# Allow localhost access
echo -e "${GREEN}Allowing localhost access to port 11434...${NC}"
sudo ufw allow in from 127.0.0.1 to any port 11434
sudo ufw allow out from any to 127.0.0.1 port 11434

# Block external access
echo -e "${GREEN}Blocking external access to port 11434...${NC}"
sudo ufw deny in to any port 11434
sudo ufw deny out to any port 11434

Scans for known LLM processes including:

ollama
llama
llamacpp
text-generation-server
python
pytorch
tensorflow

Check for any rules allowing external access

sudo ufw status verbose | grep 11434

Check what address Ollama is bound to

sudo netstat -tulpn | grep 11434

d

Verify socket binding

sudo lsof -i :11434

stops Ollama daemon service if running provides custom network port for scanning

 git clone https://github.com/aimldr/doubletap.git

cd doubletap && chmod +x doubletap.sh && ./doubletap.sh

enter a port number within 3 seconds or default :11434
requires root privileges for stopping and killing llama

y/N option for liberty with ufw and blocks external llama interaction while maintaining localhost as default
performs audit
edit to your specifications