How samugit83/redamon works
RedAmon is a self-hosted AI penetration-testing platform. Its documented main workflow moves from authorized reconnaissance through graph-backed attack-surface analysis and human-gated exploitation, then AI triage, automated code remediation, and a GitHub pull request. The graph also preserves distinct AI-surface scanning, supply-chain scanning, traffic capture, and knowledge-base capabilities. The README supports the high-level workflow; source-index references and excerpts support only the cited internal links, so other connections are omitted rather than inferred.
Main entry points: main.py Recon Controller · agentic/orchestrator.py Agent Orchestrator · neo4j_client.py Neo4j Graph · cypherfix_triage/orchestrator.py Finding Triage
Components 21
Reconnaissance
- Recon Controller
recon/main.py - Recon Modules
recon/main_recon_modules - Recon Settings
recon/project_settings.py - Recon Batching
recon/helpers/batch_groups.py - Recon Circuit Breaker
recon/helpers/circuit_breaker.py
Attack Graph
- Neo4j Graph
graph_db/neo4j_client.py - Resource Governor
graph_db/resource_governor.py
Agentic Testing
- Agent Orchestrator
agentic/orchestrator.py - Agent State
agentic/state.py - Human Approval Loop
agentic/agent_in_the_loop/broker.py - Sandbox Tools
mcp/servers - LLM Guard
agentic/llm_guard.py
Triage and Fixes
- Finding Triage
agentic/cypherfix_triage/orchestrator.py - Triage Scoring
agentic/cypherfix_triage/score_model.py - CodeFix Agent
agentic/cypherfix_codefix/orchestrator.py - Workspace Import
agentic/workspace_import/http_routes.py
Security Capabilities
- AI Surface Scan
scanners/ai_attack_surface_scan/main.py - Supply-Chain Scan
scanners/supply_chain_scan/main.py - Traffic Capture
scanners/capture_proxy/ingest_worker.py - Security Knowledge Base
services/knowledge_base/curation/data_ingestion.py
Other components
- Security Operator
Connections 25
- Security Operator → Recon Controller: starts recon
README.md - Recon Controller → Recon Modules: runs modules
recon/main.py - Recon Controller → Recon Settings: loads settings
recon/main.py - Recon Modules → Neo4j Graph: builds attack graph
README.md - Recon Controller → Resource Governor: uses resource limits
- Security Operator → Agent Orchestrator: starts testing
README.md - Agent Orchestrator → Human Approval Loop: requests approval
README.md - Agent Orchestrator → Sandbox Tools: runs tools
README.md - Agent Orchestrator → Neo4j Graph: queries attack graph
README.md - Agent Orchestrator → Agent State: tracks run state
- Security Operator → AI Surface Scan: launches AI scan
- AI Surface Scan → Neo4j Graph: writes findings
scanners/ai_attack_surface_scan/main.py - Security Operator → Supply-Chain Scan: submits scan
- Supply-Chain Scan → Neo4j Graph: writes findings
scanners/supply_chain_scan/main.py - Traffic Capture → Neo4j Graph: stores transactions
- Security Knowledge Base → Neo4j Graph: loads knowledge
services/knowledge_base/curation/data_ingestion.py - Agent Orchestrator → Finding Triage: passes findings
README.md - Finding Triage → Triage Scoring: scores findings
agentic/cypherfix_triage/orchestrator.py - Finding Triage → CodeFix Agent: hands off remediation
README.md - CodeFix Agent → Security Operator: proposes pull request
README.md - Security Operator → Workspace Import: imports project files
- Workspace Import → LLM Guard: checks requests
agentic/workspace_import/http_routes.py - Human Approval Loop → LLM Guard: checks requests
agentic/agent_in_the_loop/routes.py - Recon Controller → Recon Batching: uses batch settings
recon/main.py - Recon Controller → Recon Circuit Breaker: uses scan controls
recon/main.py