Add read-only Debug API support and update documentation

- Introduced a local read-only Debug API accessible at `http://127.0.0.1:17821` for diagnostics and agent interactions.
- Updated CLI commands to include `gpu-rent debug` for launching the Debug API.
- Enhanced documentation to reflect the new Debug API features and usage.
- Modified configuration to include `DEBUG_LOCAL_PORT` for easier customization.
- Added tests to ensure Debug API links are correctly generated in access card outputs.

docs/architecture.md

@@ -5,11 +5,13 @@
 ```
 ┌─ локальная машина (Windows / Linux) ─────────────────────────┐
 │  gpu-rent CLI                                                 │
-│    up / stop / status / doctor / hold                         │
+│    up / stop / status / doctor / hold / debug                 │
 │    tunnel / open          — SSH localhost:17801 → VM :7801    │
+│    Debug API (read-only)  — http://127.0.0.1:17821            │
 │    state <repo>/.gpu-rent/state.json                          │
 │                                                               │
 │  браузер / MCP / curl API →  http://127.0.0.1:17801           │
+│  агент / curl debug       →  http://127.0.0.1:17821/snapshot  │
 │  локальный SwarmUI        →  http://127.0.0.1:7801  (не трогаем)
 └───────────────────────────────┬──────────────────────────────┘
                                 │ SSH :22  и  OpenStack API
@@ -48,6 +50,7 @@
 | `sync_files` | SFTP `Models` / Wildcards / workflows / Output |
 | `notify` | Toast/звук при backend Idle |
 | `tunnel` | sshtunnel + Nova EXPIRED watchdog |
+| `debug_api` / `debug_checks` | localhost read-only HTTP sidecar (`:17821`) для агента |
 | `local_watchdog` | Опциональный локальный тик → stop при unclean exit |
 | `llm_runtime` / `setup_wizard` | Opt-in Ollama + `ollama-models.yaml` |
 | `idle_killer` / `hold` | systemd на VM + hold-файл |