How skygod07/rakshastra works
Rakshastra is an autonomous cybersecurity platform: analysts or connected channels submit security requests, the agent coordinates model-backed analysis and security tools, and the platform supports investigation, threat intelligence, and response. Its documented capabilities also include graph-based attribution and attack prediction, hybrid threat-intelligence RAG, UEBA, SOAR, a REST API, and x402 billing. The README supports the product-level flow; source-index references establish only the specific internal links cited below, so unsampled subsystems are shown without speculative connections.
Main entry points: tool_executor.py Tool execution · run_agent.py Agent entry · api_server.py REST API · conversation_loop.py Conversation loop
Components 23
User interfaces
- Web dashboard
web/src/App.tsx - Terminal UI
ui-tui/packages/rakshastra-ink/src/ink/components/App.tsx - CLI
rakshastra_cli/main.py - ACP server
acp_adapter/server.py
Ingress and channels
- REST API
gateway/platforms/api_server.py - Messaging channels
gateway/platforms/signal.py - Gateway daemon
gateway/run.py - x402 billing
gateway/platforms/api_server.py
Agent runtime
- Agent entry
run_agent.py - Conversation loop
agent/conversation_loop.py - Model adapters
agent/anthropic_adapter.py - Tool execution
agent/tool_executor.py - Memory manager
agent/memory_manager.py
Security intelligence
- Footprint correlation
rakshastra_core/intelligence/correlation_engine.py - Attack graph
rakshastra_core/intelligence/graph_engine.py - APT attribution
services/apt_service.py - Threat intel RAG
stage 2/rakshastra_core/intelligence/threat_intel_rag.py - Behavior analytics
rakshastra_core/intelligence - Blast radius
rakshastra_core/engines/digital_twin_engine.py
Response and evidence
- Incident response
rakshastra_core/engines/incident_response_engine.py - SOAR playbooks
rakshastra_core/engines/soar_engine.py - Evidence handling
tools/terminal_tool.py
Other components
- Analyst
Connections 26
- Analyst → Web dashboard: submits request
README.md - Analyst → Terminal UI: submits request
README.md - Analyst → CLI: submits request
README.md - Analyst → ACP server: submits request
README.md - Analyst → REST API: submits request
README.md - Messaging channels → Gateway daemon: bridges messages
README.md - Gateway daemon → Agent entry: dispatches requests
- REST API → Agent entry: dispatches requests
- Web dashboard → REST API: uses API
- Terminal UI → Gateway daemon: uses gateway
- ACP server → Agent entry: exposes agent
acp_adapter/server.py - CLI → Agent entry: starts agent
- Agent entry → Conversation loop: runs loop
agent/conversation_loop.py - Conversation loop → Model adapters: requests completion
agent/conversation_loop.py - Conversation loop → Tool execution: executes tools
- Agent entry → Memory manager: manages memory
agent/memory_manager.py - Tool execution → Footprint correlation: runs analysis
- Tool execution → Attack graph: runs analysis
- Tool execution → APT attribution: runs attribution
- Tool execution → Threat intel RAG: retrieves intelligence
- Tool execution → Behavior analytics: runs detection
- Tool execution → Blast radius: simulates impact
- Tool execution → Incident response: supports response
- Tool execution → SOAR playbooks: runs playbooks
- Tool execution → Evidence handling: uses terminal
- REST API → x402 billing: verifies payment