OpenSIN-AI Code — Python Agent Development Platform
OpenSIN-AI Code ist die Python-basierte Agent Development Platform der OpenSIN-AI Organisation. Metadata-driven Architektur für schnellen Prototyping und Porting.
Repository: OpenSIN-AI/opensin-ai-code
Umfang: 100 Dateien | 2.386 Zeilen Python-Code | 26 Subsystem-Pakete
OpenSIN-AI Agent Roadmap
- Feature spec: OpenSIN-overview/docs/opensin-ai-agent-feature-spec.md
- Comparison guide: OpenSIN-AI Agent Features
- This repo maps the Python agent development surface to the same OpenSIN-AI Agent roadmap.
Architektur
OpenSIN-AI Code ist als metadata-driven Python Workspace organisiert:
opensin-ai-code/
├── README.md # Projektdokumentation
├── LICENSE # Apache 2.0
├── src/
│ ├── main.py # CLI Entry Point (argparse)
│ ├── runtime.py # PortRuntime: Routing, Bootstrap, Turn Loops
│ ├── query_engine.py # QueryEnginePort: Message Submission, Streaming
│ ├── QueryEngine.py # QueryEngineRuntime Wrapper
│ ├── commands.py # Command Surface (207 entries)
│ ├── tools.py # Tool Surface (184 entries)
│ ├── session_store.py # Session Persistence
│ ├── transcript.py # Transcript Buffer
│ ├── port_manifest.py # Workspace Manifest
│ ├── parity_audit.py # Python vs TypeScript Coverage
│ ├── command_graph.py # Command Segmentation
│ ├── bootstrap_graph.py # Bootstrap Stages
│ ├── tool_pool.py # Tool Pool Assembly
│ ├── permissions.py # Tool Permission Context
│ ├── execution_registry.py # Mirrored Command/Tool Wrappers
│ ├── setup.py # Setup Report
│ ├── prefetch.py # Simulated Prefetches
│ ├── deferred_init.py # Trust-gated Init
│ ├── system_init.py # System Init Message Builder
│ ├── models.py # Core Dataclasses
│ ├── context.py # Port Context
│ ├── history.py # Event Log
│ ├── cost_tracker.py # Cost Tracking
│ ├── costHook.py # Cost Hook Application
│ ├── remote_runtime.py # Remote Mode Simulation
│ ├── direct_modes.py # Direct-connect Modes
│ ├── task.py # Task Stub
│ ├── tasks.py # Default Tasks
│ ├── Tool.py # Tool Definitions
│ ├── query.py # Query DTOs
│ ├── replLauncher.py # REPL Banner
│ ├── dialogLaunchers.py # Dialog Launchers
│ ├── ink.py # Markdown Rendering
│ ├── interactiveHelpers.py # Utility Helpers
│ ├── projectOnboardingState.py # Onboarding State
│ └── [26 subdirectories] # Placeholder packages loading JSON metadata
└── tests/
└── test_porting_workspace.pyCore Components
CLI Entry Point (main.py)
Argparse-basierter Router mit 25+ Subcommands:
summary— Porting-Zusammenfassungmanifest— Workspace Manifestparity-audit— Coverage-Vergleichroute,bootstrap,turn-loop— Runtime-Simulationremote-mode,ssh-mode— Remote-Modi
QueryEngine (query_engine.py)
QueryEnginePort Klasse:
- Message Submission
- Turn Loop Execution
- Streaming Support
- Session Persistence
- Structured Output
- Context Compacting
PortRuntime (runtime.py)
PortRuntime Klasse:
- Prompt Routing zu Commands/Tools
- Session Bootstrap
- Multi-Turn Loop Execution
- Permission Denial Inference
Reference Data
JSON Snapshots in reference_data/:
archive_surface_snapshot.json— 1902 TS-Dateien, 207 Commands, 184 Toolscommands_snapshot.json— 207 Command Entriestools_snapshot.json— 184 Tool Entriessubsystems/*.json— 29 Subsystem Metadata Files
Subsystem Packages
| Subsystem | Beschreibung |
|---|---|
assistant/ | Assistant Module |
bootstrap/ | Bootstrap Module |
bridge/ | Bridge Module |
buddy/ | Buddy Companion |
cli/ | CLI Module |
components/ | UI Components |
constants/ | Constants |
coordinator/ | Task Coordinator |
entrypoints/ | Entry Points |
hooks/ | Hook System |
keybindings/ | Keybindings |
memdir/ | Memory Directory |
migrations/ | Migrations |
moreright/ | MoreRight Module |
native_ts/ | Native TypeScript Bridge |
outputStyles/ | Output Styles |
plugins/ | Plugin System |
remote/ | Remote Module |
schemas/ | Data Schemas |
screens/ | Screen Definitions |
server/ | Server Module |
services/ | Services |
skills/ | Skills System |
state/ | State Management |
types/ | Type Definitions |
upstreamproxy/ | Upstream Proxy |
utils/ | Utilities |
vim/ | Vim Integration |
voice/ | Voice Module |
Schnellstart
bash
# Repository klonen
git clone https://github.com/OpenSIN-AI/opensin-ai-code.git
cd opensin-ai-code
# Porting-Zusammenfassung
python3 -m src.main summary
# Workspace Manifest
python3 -m src.main manifest
# Subsysteme auflisten
python3 -m src.main subsystems --limit 16
# Tests ausführen
python3 -m unittest discover -s tests -vZuletzt aktualisiert: 2026-04-07 | OpenSIN-AI