@startuml
title Artifact Editor - Threat Model

actor "Malicious User" as attacker
actor "Legitimate User" as user

package "Artifact Editor" {
  component "CLI Interface" as cli
  component "Voice Input" as voice
  component "AI Assistant" as ai
  component "File Operations" as files
  
  package "Renderers" {
    component "PlantUML" as plantuml
    component "OpenSCAD" as openscad
    component "Mermaid" as mermaid
    component "SVG" as svg
  }
}

package "External" {
  component "PlantUML JAR" as plantumljava
  component "OpenSCAD Binary" as openscadbinary
  component "AI Provider" as aiprovider
  database "Filesystem" as fs
}

package "Parent App" as parent

user --> cli : normal usage
attacker --> cli : malicious input

cli --> voice : audio data
cli --> ai : edit requests
cli --> files : save/load
cli ..> parent : stdout protocol

voice --> aiprovider : transcription
ai --> aiprovider : generation

plantuml --> plantumljava : execute
openscad --> openscadbinary : execute

files --> fs : read/write

note right of attacker
  **Attack Vectors:**
  1. Path traversal in --output
  2. Command injection via content
  3. Prompt injection via voice
  4. Malicious AI-generated code
  5. Protocol injection in stdout
  6. Resource exhaustion
  7. Compromised dependencies
end note

note left of plantuml
  **High Risk:**
  External code execution
  File system access
  Network capabilities
end note

note bottom of files
  **Critical:**
  Validate all paths
  Restrict to artifact dir
  Check for traversal
end note

@enduml