Overview
This page summarizes features and changes projected for upcoming product releases. This list is a projection and does not guarantee that any of the features will be implemented as planned.
Future Releases
Version 6.0.0: LLM/AI Agents
Goal: Provide LLM integration that enables one or more AI agents to work together to command and control a SpecterInsight C2 server to achieve operator specified objectives.
Features
- Agent Profiles: Provide custom agent initialization prompts that define how the agent can interact with the provided SpecterInsight APIs.
- Integration Functions: Documentation on which functions are available for the AI agents to call.
- Agent Sessions: An interactive chat session where an operator tasks one or more AI agents to accomplish specified tasks. Agents will then task implants and review results until the desired end state is achieved.
- Start, stop, and pause sessions
- Cancel tasks (e.g. LLM response)
- Option for human-in-the-loop or fully autonomous execution.
- Support for multiple models and families:
- ChatGPT
- Llama
- Local LLM
Future SpecterScripts
- Persistence via Ghost Task and PowerShell Cradle
- Lateral Movement via Ghost Task and PowerShell Cradle
- Get Drivers: List drivers and their elevation for detecting obfuscated EDR deployment.
- Unload Filter Driver: disable monitoring and protection tools.
- Get Group Policy Passwords