release: version 0.9.9 🚀

Containerfile

@@ -12,7 +12,10 @@ WORKDIR /app
 
 # Install system dependencies
 RUN apt-get update && apt-get install -y --no-install-recommends \
+    ca-certificates \
     curl \
+    git \
+    && update-ca-certificates \
     && rm -rf /var/lib/apt/lists/*
 
 # Install dependencies

HISTORY.md

@@ -5,11 +5,124 @@ Changelog
 (unreleased)
 ------------
 
+Fix
+~~~
+- Add missing git binary, refs NOISSUE. [Simon Diesenreiter]
+
+
+0.9.8 (2026-04-11)
+------------------
+
+Fix
+~~~
+- More file change fixes, refs NOISSUE. [Simon Diesenreiter]
+
+Other
+~~~~~
+
+
+0.9.7 (2026-04-11)
+------------------
+
+Fix
+~~~
+- More file generation improvements, refs NOISSUE. [Simon Diesenreiter]
+
+Other
+~~~~~
+
+
+0.9.6 (2026-04-11)
+------------------
+
+Fix
+~~~
+- Repo onboarding fix, refs NOISSUE. [Simon Diesenreiter]
+
+Other
+~~~~~
+
+
+0.9.5 (2026-04-11)
+------------------
+
+Fix
+~~~
+- Better code generation, refs NOISSUE. [Simon Diesenreiter]
+
+Other
+~~~~~
+
+
+0.9.4 (2026-04-11)
+------------------
+
+Fix
+~~~
+- Add commit retry, refs NOISSUE. [Simon Diesenreiter]
+
+Other
+~~~~~
+
+
+0.9.3 (2026-04-11)
+------------------
+
+Fix
+~~~
+- Better home assistant integration, refs NOISSUE. [Simon Diesenreiter]
+
+Other
+~~~~~
+
+
+0.9.2 (2026-04-11)
+------------------
+
+Fix
+~~~
+- UI improvements and prompt hardening, refs NOISSUE. [Simon
+  Diesenreiter]
+
+Other
+~~~~~
+
+
+0.9.1 (2026-04-11)
+------------------
+
+Fix
+~~~
+- Better repo name generation, refs NOISSUE. [Simon Diesenreiter]
+
+Other
+~~~~~
+
+
+0.9.0 (2026-04-11)
+------------------
+- Feat: editable guardrails, refs NOISSUE. [Simon Diesenreiter]
+
+
+0.8.0 (2026-04-11)
+------------------
+- Feat: better dashboard reloading mechanism, refs NOISSUE. [Simon
+  Diesenreiter]
+- Feat: add explicit workflow steps and guardrail prompts, refs NOISSUE.
+  [Simon Diesenreiter]
+
+
+0.7.1 (2026-04-11)
+------------------
+
 Fix
 ~~~
 - Add additional deletion confirmation, refs NOISSUE. [Simon
   Diesenreiter]
 
+Other
+~~~~~
+
 
 0.7.0 (2026-04-10)
 ------------------

README.md

@@ -48,6 +48,7 @@ OLLAMA_URL=http://localhost:11434
 OLLAMA_MODEL=llama3
 
 # Gitea
+# Host-only values such as git.disi.dev are normalized to https://git.disi.dev.
 GITEA_URL=https://gitea.yourserver.com
 GITEA_TOKEN=your_gitea_api_token
 GITEA_OWNER=ai-software-factory
@@ -69,6 +70,12 @@ N8N_WEBHOOK_URL=http://n8n.yourserver.com/webhook/telegram
 # Telegram
 TELEGRAM_BOT_TOKEN=your_telegram_bot_token
 TELEGRAM_CHAT_ID=your_chat_id
+
+# Optional: Home Assistant integration.
+# Only the base URL and token are required in the environment.
+# Entity ids, thresholds, and queue behavior can be configured from the dashboard System tab and are stored in the database.
+HOME_ASSISTANT_URL=http://homeassistant.local:8123
+HOME_ASSISTANT_TOKEN=your_home_assistant_long_lived_token
 ```
 
 ### Build and Run
@@ -93,6 +100,7 @@ docker-compose up -d
 
    The backend now interprets free-form Telegram text with Ollama before generation.
    If `TELEGRAM_CHAT_ID` is set, the Telegram-trigger workflow only reacts to messages from that specific chat.
+   If queueing is enabled from the dashboard System tab, Telegram prompts are stored in a durable queue and processed only when the configured Home Assistant battery and surplus thresholds are satisfied, unless you force processing via `/queue/process` or send `process_now=true`.
 
 2. **Monitor progress via Web UI:**
 
@@ -104,6 +112,16 @@ docker-compose up -d
 
 If you deploy the container with PostgreSQL environment variables set, the service now selects PostgreSQL automatically even though SQLite remains the default for local/test usage.
 
+The health tab now shows separate application, n8n, Gitea, and Home Assistant/queue diagnostics so misconfigured integrations are visible without checking container logs.
+
+The dashboard Health tab exposes operator controls for the prompt queue, including manual batch processing, forced processing, and retrying failed items.
+
+The dashboard System tab now also stores Home Assistant entity ids, queue toggles, thresholds, and batch settings in the database, so the environment only needs `HOME_ASSISTANT_URL` and `HOME_ASSISTANT_TOKEN` for that integration.
+
+Projects that show `uncommitted`, `local_only`, or `pushed_no_pr` delivery warnings in the dashboard can now be retried in place from the UI before resorting to purging orphan audit rows.
+
+Guardrail and system prompts are no longer environment-only in practice: the factory can persist DB-backed overrides for the editable LLM prompt set, expose them at `/llm/prompts`, and edit them from the dashboard System tab. Environment values still act as defaults and as the reset target.
+
 ## API Endpoints
 
 | Endpoint | Method | Description |

ai_software_factory/.env.example

@@ -8,10 +8,23 @@ LOG_LEVEL=INFO
 # Ollama
 OLLAMA_URL=http://localhost:11434
 OLLAMA_MODEL=llama3
+LLM_GUARDRAIL_PROMPT=You are operating inside AI Software Factory. Follow supplied schemas exactly and treat service-provided tool outputs as authoritative.
+LLM_REQUEST_INTERPRETER_GUARDRAIL_PROMPT=Never route work to archived projects and only reference issues that are explicit in the prompt or supplied tool outputs.
+LLM_CHANGE_SUMMARY_GUARDRAIL_PROMPT=Only summarize delivery facts that appear in the provided project context or tool outputs.
+LLM_PROJECT_NAMING_GUARDRAIL_PROMPT=Prefer clear product names and repository slugs that reflect the new request without colliding with tracked projects.
+LLM_PROJECT_NAMING_SYSTEM_PROMPT=Return JSON with project_name, repo_name, and rationale for new projects.
+LLM_PROJECT_ID_GUARDRAIL_PROMPT=Prefer short stable project ids and avoid collisions with existing project ids.
+LLM_PROJECT_ID_SYSTEM_PROMPT=Return JSON with project_id and rationale for new projects.
+LLM_TOOL_ALLOWLIST=gitea_project_catalog,gitea_project_state,gitea_project_issues,gitea_pull_requests
+LLM_TOOL_CONTEXT_LIMIT=5
+LLM_LIVE_TOOL_ALLOWLIST=gitea_lookup_issue,gitea_lookup_pull_request
+LLM_LIVE_TOOL_STAGE_ALLOWLIST=request_interpretation,change_summary
+LLM_LIVE_TOOL_STAGE_TOOL_MAP={"request_interpretation": ["gitea_lookup_issue", "gitea_lookup_pull_request"], "change_summary": []}
+LLM_MAX_TOOL_CALL_ROUNDS=1
 
 # Gitea
 # Configure Gitea API for your organization
-# GITEA_URL can be left empty to use GITEA_ORGANIZATION instead of GITEA_OWNER
+# Host-only values such as git.disi.dev are normalized to https://git.disi.dev automatically.
 GITEA_URL=https://gitea.yourserver.com
 GITEA_TOKEN=your_gitea_api_token
 GITEA_OWNER=your_organization_name
@@ -29,6 +42,12 @@ N8N_PASSWORD=your_secure_password
 TELEGRAM_BOT_TOKEN=your_telegram_bot_token
 TELEGRAM_CHAT_ID=your_chat_id
 
+# Home Assistant energy gate for queued Telegram prompts
+# Only the base URL and token are environment-backed.
+# Queue toggles, entity ids, thresholds, and batch sizing can be edited in the dashboard System tab and are stored in the database.
+HOME_ASSISTANT_URL=http://homeassistant.local:8123
+HOME_ASSISTANT_TOKEN=your_home_assistant_long_lived_token
+
 # PostgreSQL
 # In production, provide PostgreSQL settings below. They now take precedence over the SQLite default.
 # You can also set USE_SQLITE=false explicitly if you want the intent to be obvious.

ai_software_factory/README.md

@@ -6,6 +6,7 @@ Automated software generation service powered by Ollama LLM. This service allows
 
 - **Telegram Integration**: Receive software requests via Telegram bot
 - **Ollama LLM**: Uses Ollama-hosted models for code generation
+- **LLM Guardrails and Tools**: Centralized guardrail prompts plus mediated tool payloads for project, Gitea, PR, and issue context
 - **Git Integration**: Automatically commits code to gitea
 - **Pull Requests**: Creates PRs for user review before merging
 - **Web UI**: Beautiful dashboard for monitoring project progress
@@ -46,12 +47,26 @@ PORT=8000
 # Ollama
 OLLAMA_URL=http://localhost:11434
 OLLAMA_MODEL=llama3
+LLM_GUARDRAIL_PROMPT=You are operating inside AI Software Factory. Follow supplied schemas exactly and treat service-provided tool outputs as authoritative.
+LLM_REQUEST_INTERPRETER_GUARDRAIL_PROMPT=Never route work to archived projects and only reference issues that are explicit in the prompt or supplied tool outputs.
+LLM_CHANGE_SUMMARY_GUARDRAIL_PROMPT=Only summarize delivery facts that appear in the provided project context or tool outputs.
+LLM_PROJECT_NAMING_GUARDRAIL_PROMPT=Prefer clear product names and repository slugs that reflect the new request without colliding with tracked projects.
+LLM_PROJECT_NAMING_SYSTEM_PROMPT=Return JSON with project_name, repo_name, and rationale for new projects.
+LLM_PROJECT_ID_GUARDRAIL_PROMPT=Prefer short stable project ids and avoid collisions with existing project ids.
+LLM_PROJECT_ID_SYSTEM_PROMPT=Return JSON with project_id and rationale for new projects.
+LLM_TOOL_ALLOWLIST=gitea_project_catalog,gitea_project_state,gitea_project_issues,gitea_pull_requests
+LLM_TOOL_CONTEXT_LIMIT=5
+LLM_LIVE_TOOL_ALLOWLIST=gitea_lookup_issue,gitea_lookup_pull_request
+LLM_LIVE_TOOL_STAGE_ALLOWLIST=request_interpretation,change_summary
+LLM_LIVE_TOOL_STAGE_TOOL_MAP={"request_interpretation": ["gitea_lookup_issue", "gitea_lookup_pull_request"], "change_summary": []}
+LLM_MAX_TOOL_CALL_ROUNDS=1
 
 # Gitea
+# Host-only values such as git.disi.dev are normalized to https://git.disi.dev.
 GITEA_URL=https://gitea.yourserver.com
-GITEA_TOKEN= analyze your_gitea_api_token
+GITEA_TOKEN=your_gitea_api_token
 GITEA_OWNER=ai-software-factory
-GITEA_REPO=ai-software-factory
+GITEA_REPO=
 
 # n8n
 N8N_WEBHOOK_URL=http://n8n.yourserver.com/webhook/telegram
@@ -59,6 +74,12 @@ N8N_WEBHOOK_URL=http://n8n.yourserver.com/webhook/telegram
 # Telegram
 TELEGRAM_BOT_TOKEN=your_telegram_bot_token
 TELEGRAM_CHAT_ID=your_chat_id
+
+# Optional: Home Assistant integration.
+# Only the base URL and token are required in the environment.
+# Entity ids, thresholds, and queue behavior can be configured from the dashboard System tab and are stored in the database.
+HOME_ASSISTANT_URL=http://homeassistant.local:8123
+HOME_ASSISTANT_TOKEN=your_home_assistant_long_lived_token
 ```
 
 ### Build and Run
@@ -81,6 +102,10 @@ docker-compose up -d
    Features: user authentication, task CRUD, notifications
    ```
 
+   If queueing is enabled from the dashboard System tab, Telegram prompts are queued durably and processed only when Home Assistant reports the configured battery and surplus thresholds. Operators can override the gate via `/queue/process` or by sending `process_now=true` to `/generate/text`.
+
+The dashboard System tab stores Home Assistant entity ids, queue toggles, thresholds, and batch settings in the database, so the environment only needs `HOME_ASSISTANT_URL` and `HOME_ASSISTANT_TOKEN` for that integration.
+
 2. **Monitor progress via Web UI:**
 
    Open `http://yourserver:8000` to see real-time progress
@@ -99,6 +124,39 @@ docker-compose up -d
 | `/status/{project_id}` | GET | Get project status |
 | `/projects` | GET | List all projects |
 
+## LLM Guardrails and Tool Access
+
+External LLM calls are now routed through a centralized client that applies:
+
+- A global guardrail prompt for every outbound model request
+- Stage-specific guardrails for request interpretation and change summaries
+- Service-mediated tool outputs that expose tracked Gitea/project state without giving the model raw credentials
+
+Current mediated tools include:
+
+- `gitea_project_catalog`: active tracked projects and repository mappings
+- `gitea_project_state`: current repository, PR, and linked-issue state for the project in scope
+- `gitea_project_issues`: tracked open issues for the relevant repository
+- `gitea_pull_requests`: tracked pull requests for the relevant repository
+
+The service also supports a bounded live tool-call loop for selected lookups. When enabled, the model may request one live call such as `gitea_lookup_issue` or `gitea_lookup_pull_request`, the service executes it against Gitea, and the final model response is generated from the returned result. This remains mediated by the service, so the model never receives raw credentials.
+
+Live tool access is stage-aware. `LLM_LIVE_TOOL_ALLOWLIST` controls which live tools exist globally, while `LLM_LIVE_TOOL_STAGE_ALLOWLIST` controls which LLM stages may use them. If you need per-stage subsets, `LLM_LIVE_TOOL_STAGE_TOOL_MAP` accepts a JSON object mapping each stage to the exact tools it may use. For example, you can allow issue and PR lookups during `request_interpretation` while keeping `change_summary` fully read-only.
+
+When the interpreter decides a prompt starts a new project, the service can run a dedicated `project_naming` LLM stage before generation. `LLM_PROJECT_NAMING_SYSTEM_PROMPT` and `LLM_PROJECT_NAMING_GUARDRAIL_PROMPT` let you steer how project titles and repository slugs are chosen. The interpreter now checks tracked project repositories plus live Gitea repository names when available, so if the model suggests a colliding repo slug the service will automatically move to the next available slug.
+
+New project creation can also run a dedicated `project_id_naming` stage. `LLM_PROJECT_ID_SYSTEM_PROMPT` and `LLM_PROJECT_ID_GUARDRAIL_PROMPT` control how stable project ids are chosen, and the service will append deterministic numeric suffixes when an id is already taken instead of always falling back to a random UUID-based id.
+
+Runtime visibility for the active guardrails, mediated tools, live tools, and model configuration is available at `/llm/runtime` and in the dashboard System tab.
+
+Operational visibility for the Gitea integration, Home Assistant energy gate, and queued prompt counts is available in the dashboard Health tab, plus `/gitea/health`, `/home-assistant/health`, and `/queue`.
+
+The dashboard Health tab also includes operator controls for manually processing queued Telegram prompts, force-processing them when needed, and retrying failed items.
+
+Editable guardrail and system prompts are persisted in the database as overrides on top of the environment defaults. The current merged values are available at `/llm/prompts`, and the dashboard System tab can edit or reset them without restarting the service.
+
+These tool payloads are appended to the model prompt as authoritative JSON generated by the service, so the LLM can reason over live project and Gitea context while remaining constrained by the configured guardrails.
+
 ## Development
 
 ### Makefile Targets

ai_software_factory/VERSION

@@ -1 +1 @@
-0.7.1
+0.9.9

ai_software_factory/agents/change_summary.py

@@ -4,8 +4,10 @@ from __future__ import annotations
 
 try:
     from ..config import settings
+    from .llm_service import LLMServiceClient
 except ImportError:
     from config import settings
+    from agents.llm_service import LLMServiceClient
 
 
 class ChangeSummaryGenerator:
@@ -14,6 +16,7 @@ class ChangeSummaryGenerator:
     def __init__(self, ollama_url: str | None = None, model: str | None = None):
         self.ollama_url = (ollama_url or settings.ollama_url).rstrip('/')
         self.model = model or settings.OLLAMA_MODEL
+        self.llm_client = LLMServiceClient(ollama_url=self.ollama_url, model=self.model)
 
     async def summarize(self, context: dict) -> str:
         """Summarize project changes with Ollama, or fall back to a deterministic overview."""
@@ -28,40 +31,24 @@ class ChangeSummaryGenerator:
             'Write 3 to 5 sentences. Mention the application goal, main delivered pieces, '
             'technical direction, and what the user should expect next. Avoid markdown bullets.'
         )
-        try:
-            import aiohttp
-
-            async with aiohttp.ClientSession() as session:
-                async with session.post(
-                    f'{self.ollama_url}/api/chat',
-                    json={
-                        'model': self.model,
-                        'stream': False,
-                        'messages': [
-                            {
-                                'role': 'system',
-                                'content': system_prompt,
+        content, trace = await self.llm_client.chat_with_trace(
+            stage='change_summary',
+            system_prompt=system_prompt,
+            user_prompt=prompt,
+            tool_context_input={
+                'project_id': context.get('project_id'),
+                'project_name': context.get('name'),
+                'repository': context.get('repository'),
+                'repository_url': context.get('repository_url'),
+                'pull_request': context.get('pull_request'),
+                'pull_request_url': context.get('pull_request_url'),
+                'pull_request_state': context.get('pull_request_state'),
+                'related_issue': context.get('related_issue'),
+                'issues': [context.get('related_issue')] if context.get('related_issue') else [],
             },
-                            {'role': 'user', 'content': prompt},
-                        ],
-                    },
-                ) as resp:
-                    payload = await resp.json()
-                    if 200 <= resp.status < 300:
-                        content = payload.get('message', {}).get('content', '').strip()
+        )
         if content:
-                            return content, {
-                                'stage': 'change_summary',
-                                'provider': 'ollama',
-                                'model': self.model,
-                                'system_prompt': system_prompt,
-                                'user_prompt': prompt,
-                                'assistant_response': content,
-                                'raw_response': payload,
-                                'fallback_used': False,
-                            }
-        except Exception:
-            pass
+            return content.strip(), trace
 
         fallback = self._fallback(context)
         return fallback, {
@@ -71,7 +58,9 @@ class ChangeSummaryGenerator:
             'system_prompt': system_prompt,
             'user_prompt': prompt,
             'assistant_response': fallback,
-            'raw_response': {'fallback': 'deterministic'},
+            'raw_response': {'fallback': 'deterministic', 'llm_trace': trace.get('raw_response') if isinstance(trace, dict) else None},
+            'guardrails': trace.get('guardrails') if isinstance(trace, dict) else [],
+            'tool_context': trace.get('tool_context') if isinstance(trace, dict) else [],
             'fallback_used': True,
         }
 

ai_software_factory/agents/gitea.py

@@ -4,6 +4,20 @@ import os
 import urllib.error
 import urllib.request
 import json
+from urllib.parse import urlparse
+
+
+def _normalize_base_url(base_url: str) -> str:
+    """Normalize host-only service addresses into valid absolute URLs."""
+    normalized = (base_url or '').strip().rstrip('/')
+    if not normalized:
+        return ''
+    if '://' not in normalized:
+        normalized = f'https://{normalized}'
+    parsed = urlparse(normalized)
+    if not parsed.scheme or not parsed.netloc:
+        return ''
+    return normalized
 
 
 class GiteaAPI:
@@ -11,7 +25,7 @@ class GiteaAPI:
 
     def __init__(self, token: str, base_url: str, owner: str | None = None, repo: str | None = None):
         self.token = token
-        self.base_url = base_url.rstrip("/")
+        self.base_url = _normalize_base_url(base_url)
         self.owner = owner
         self.repo = repo
         self.headers = {
@@ -26,7 +40,7 @@ class GiteaAPI:
         owner = os.getenv("GITEA_OWNER", "ai-test")
         repo = os.getenv("GITEA_REPO", "")
         return {
-            "base_url": base_url.rstrip("/"),
+            "base_url": _normalize_base_url(base_url),
             "token": token,
             "owner": owner,
             "repo": repo,
@@ -96,16 +110,16 @@ class GiteaAPI:
 
     def _request_sync(self, method: str, path: str, payload: dict | None = None) -> dict:
         """Perform a synchronous Gitea API request."""
+        try:
+            if not self.base_url:
+                return {'error': 'Gitea base URL is not configured or is invalid'}
             request = urllib.request.Request(
                 self._api_url(path),
                 headers=self.get_auth_headers(),
                 method=method.upper(),
             )
-        data = None
             if payload is not None:
-            data = json.dumps(payload).encode('utf-8')
-            request.data = data
-        try:
+                request.data = json.dumps(payload).encode('utf-8')
             with urllib.request.urlopen(request) as response:
                 body = response.read().decode('utf-8')
                 return json.loads(body) if body else {}
@@ -182,6 +196,10 @@ class GiteaAPI:
         """Get the user associated with the configured token."""
         return await self._request("GET", "user")
 
+    def get_current_user_sync(self) -> dict:
+        """Synchronously get the user associated with the configured token."""
+        return self._request_sync("GET", "user")
+
     async def create_branch(self, branch: str, base: str = "main", owner: str | None = None, repo: str | None = None):
         """Create a new branch."""
         _owner = owner or self.owner
@@ -212,6 +230,26 @@ class GiteaAPI:
         }
         return await self._request("POST", f"repos/{_owner}/{_repo}/pulls", payload)
 
+    def create_pull_request_sync(
+        self,
+        title: str,
+        body: str,
+        owner: str,
+        repo: str,
+        base: str = "main",
+        head: str | None = None,
+    ) -> dict:
+        """Synchronously create a pull request."""
+        _owner = owner or self.owner
+        _repo = repo or self.repo
+        payload = {
+            "title": title,
+            "body": body,
+            "base": base,
+            "head": head or f"{_owner}-{_repo}-ai-gen-{hash(title) % 10000}",
+        }
+        return self._request_sync("POST", f"repos/{_owner}/{_repo}/pulls", payload)
+
     async def list_pull_requests(
         self,
         owner: str | None = None,
@@ -384,3 +422,13 @@ class GiteaAPI:
             return {"error": "Repository name required for org operations"}
 
         return await self._request("GET", f"repos/{_owner}/{_repo}")
+
+    def get_repo_info_sync(self, owner: str | None = None, repo: str | None = None) -> dict:
+        """Synchronously get repository information."""
+        _owner = owner or self.owner
+        _repo = repo or self.repo
+
+        if not _repo:
+            return {"error": "Repository name required for org operations"}
+
+        return self._request_sync("GET", f"repos/{_owner}/{_repo}")

ai_software_factory/agents/home_assistant.py

@@ -0,0 +1,162 @@
+"""Home Assistant integration for energy-gated queue processing."""
+
+from __future__ import annotations
+
+try:
+    from ..config import settings
+except ImportError:
+    from config import settings
+
+
+class HomeAssistantAgent:
+    """Query Home Assistant for queue-processing eligibility and health."""
+
+    def __init__(self, base_url: str | None = None, token: str | None = None):
+        self.base_url = (base_url or settings.home_assistant_url).rstrip('/')
+        self.token = token or settings.home_assistant_token
+
+    def _headers(self) -> dict[str, str]:
+        return {
+            'Authorization': f'Bearer {self.token}',
+            'Content-Type': 'application/json',
+        }
+
+    def _state_url(self, entity_id: str) -> str:
+        return f'{self.base_url}/api/states/{entity_id}'
+
+    async def _get_state(self, entity_id: str) -> dict:
+        if not self.base_url:
+            return {'error': 'Home Assistant URL is not configured'}
+        if not self.token:
+            return {'error': 'Home Assistant token is not configured'}
+        if not entity_id:
+            return {'error': 'Home Assistant entity id is not configured'}
+        try:
+            import aiohttp
+
+            async with aiohttp.ClientSession() as session:
+                async with session.get(self._state_url(entity_id), headers=self._headers()) as resp:
+                    payload = await resp.json(content_type=None)
+                    if 200 <= resp.status < 300:
+                        return payload if isinstance(payload, dict) else {'value': payload}
+                    return {'error': payload, 'status_code': resp.status}
+        except Exception as exc:
+            return {'error': str(exc)}
+
+    def _get_state_sync(self, entity_id: str) -> dict:
+        if not self.base_url:
+            return {'error': 'Home Assistant URL is not configured'}
+        if not self.token:
+            return {'error': 'Home Assistant token is not configured'}
+        if not entity_id:
+            return {'error': 'Home Assistant entity id is not configured'}
+        try:
+            import json
+            import urllib.error
+            import urllib.request
+
+            request = urllib.request.Request(self._state_url(entity_id), headers=self._headers(), method='GET')
+            with urllib.request.urlopen(request) as response:
+                body = response.read().decode('utf-8')
+                return json.loads(body) if body else {}
+        except urllib.error.HTTPError as exc:
+            try:
+                body = exc.read().decode('utf-8')
+            except Exception:
+                body = str(exc)
+            return {'error': body, 'status_code': exc.code}
+        except Exception as exc:
+            return {'error': str(exc)}
+
+    @staticmethod
+    def _coerce_float(payload: dict) -> float | None:
+        raw = payload.get('state') if isinstance(payload, dict) else None
+        try:
+            return float(raw)
+        except Exception:
+            return None
+
+    async def queue_gate_status(self, force: bool = False) -> dict:
+        """Return whether queued prompts may be processed now."""
+        if force or settings.prompt_queue_force_process:
+            return {
+                'status': 'success',
+                'allowed': True,
+                'forced': True,
+                'reason': 'Queue override is enabled',
+            }
+        battery = await self._get_state(settings.home_assistant_battery_entity_id)
+        surplus = await self._get_state(settings.home_assistant_surplus_entity_id)
+        battery_value = self._coerce_float(battery)
+        surplus_value = self._coerce_float(surplus)
+        checks = []
+        if battery.get('error'):
+            checks.append({'name': 'battery', 'ok': False, 'message': str(battery.get('error')), 'entity_id': settings.home_assistant_battery_entity_id})
+        else:
+            checks.append({'name': 'battery', 'ok': battery_value is not None and battery_value >= settings.home_assistant_battery_full_threshold, 'message': f'{battery_value}%', 'entity_id': settings.home_assistant_battery_entity_id})
+        if surplus.get('error'):
+            checks.append({'name': 'surplus', 'ok': False, 'message': str(surplus.get('error')), 'entity_id': settings.home_assistant_surplus_entity_id})
+        else:
+            checks.append({'name': 'surplus', 'ok': surplus_value is not None and surplus_value >= settings.home_assistant_surplus_threshold_watts, 'message': f'{surplus_value} W', 'entity_id': settings.home_assistant_surplus_entity_id})
+        allowed = all(check['ok'] for check in checks)
+        return {
+            'status': 'success' if allowed else 'blocked',
+            'allowed': allowed,
+            'forced': False,
+            'checks': checks,
+            'battery_level': battery_value,
+            'surplus_watts': surplus_value,
+            'thresholds': {
+                'battery_full_percent': settings.home_assistant_battery_full_threshold,
+                'surplus_watts': settings.home_assistant_surplus_threshold_watts,
+            },
+            'reason': 'Energy gate open' if allowed else 'Battery or surplus threshold not met',
+        }
+
+    def health_check_sync(self) -> dict:
+        """Return current Home Assistant connectivity and queue gate diagnostics."""
+        if not self.base_url:
+            return {
+                'status': 'error',
+                'message': 'Home Assistant URL is not configured.',
+                'base_url': '',
+                'configured': False,
+                'checks': [],
+            }
+        if not self.token:
+            return {
+                'status': 'error',
+                'message': 'Home Assistant token is not configured.',
+                'base_url': self.base_url,
+                'configured': False,
+                'checks': [],
+            }
+        battery = self._get_state_sync(settings.home_assistant_battery_entity_id)
+        surplus = self._get_state_sync(settings.home_assistant_surplus_entity_id)
+        checks = []
+        for name, entity_id, payload in (
+            ('battery', settings.home_assistant_battery_entity_id, battery),
+            ('surplus', settings.home_assistant_surplus_entity_id, surplus),
+        ):
+            checks.append(
+                {
+                    'name': name,
+                    'entity_id': entity_id,
+                    'ok': not bool(payload.get('error')),
+                    'message': str(payload.get('error') or payload.get('state') or 'ok'),
+                    'status_code': payload.get('status_code'),
+                    'url': self._state_url(entity_id) if entity_id else self.base_url,
+                }
+            )
+        return {
+            'status': 'success' if all(check['ok'] for check in checks) else 'error',
+            'message': 'Home Assistant connectivity is healthy.' if all(check['ok'] for check in checks) else 'Home Assistant checks failed.',
+            'base_url': self.base_url,
+            'configured': True,
+            'checks': checks,
+            'queue_gate': {
+                'battery_full_percent': settings.home_assistant_battery_full_threshold,
+                'surplus_watts': settings.home_assistant_surplus_threshold_watts,
+                'force_process': settings.prompt_queue_force_process,
+            },
+        }

ai_software_factory/agents/llm_service.py

@@ -0,0 +1,394 @@
+"""Centralized LLM client with guardrails and mediated tool context."""
+
+from __future__ import annotations
+
+import json
+
+try:
+    from .gitea import GiteaAPI
+except ImportError:
+    from gitea import GiteaAPI
+
+try:
+    from ..config import settings
+except ImportError:
+    from config import settings
+
+
+class LLMToolbox:
+    """Build named tool payloads that can be shared with external LLM providers."""
+
+    SUPPORTED_LIVE_TOOL_STAGES = ('request_interpretation', 'change_summary', 'generation_plan', 'project_naming', 'project_id_naming')
+
+    def build_tool_context(self, stage: str, context: dict | None = None) -> list[dict]:
+        """Return the mediated tool payloads allowed for this LLM request."""
+        context = context or {}
+        allowed = set(settings.llm_tool_allowlist)
+        limit = settings.llm_tool_context_limit
+        tool_context: list[dict] = []
+
+        if 'gitea_project_catalog' in allowed:
+            projects = context.get('projects') or []
+            if projects:
+                tool_context.append(
+                    {
+                        'name': 'gitea_project_catalog',
+                        'description': 'Tracked active projects and their repository mappings inside the factory.',
+                        'payload': projects[:limit],
+                    }
+                )
+
+        if 'gitea_project_state' in allowed:
+            state_payload = {
+                'project_id': context.get('project_id'),
+                'project_name': context.get('project_name') or context.get('name'),
+                'repository': context.get('repository'),
+                'repository_url': context.get('repository_url'),
+                'pull_request': context.get('pull_request'),
+                'pull_request_url': context.get('pull_request_url'),
+                'pull_request_state': context.get('pull_request_state'),
+                'related_issue': context.get('related_issue'),
+            }
+            if any(value for value in state_payload.values()):
+                tool_context.append(
+                    {
+                        'name': 'gitea_project_state',
+                        'description': 'Current repository and pull-request state for the project being discussed.',
+                        'payload': state_payload,
+                    }
+                )
+
+        if 'gitea_project_issues' in allowed:
+            issues = context.get('open_issues') or context.get('issues') or []
+            if issues:
+                tool_context.append(
+                    {
+                        'name': 'gitea_project_issues',
+                        'description': 'Open tracked Gitea issues for the relevant project repository.',
+                        'payload': issues[:limit],
+                    }
+                )
+
+        if 'gitea_pull_requests' in allowed:
+            pull_requests = context.get('pull_requests') or []
+            if pull_requests:
+                tool_context.append(
+                    {
+                        'name': 'gitea_pull_requests',
+                        'description': 'Tracked pull requests associated with the relevant project repository.',
+                        'payload': pull_requests[:limit],
+                    }
+                )
+
+        return tool_context
+
+    def build_live_tool_specs(self, stage: str, context: dict | None = None) -> list[dict]:
+        """Return live tool-call specs that the model may request explicitly."""
+        _context = context or {}
+        specs = []
+        allowed = set(settings.llm_live_tools_for_stage(stage))
+        if 'gitea_lookup_issue' in allowed:
+            specs.append(
+                {
+                    'name': 'gitea_lookup_issue',
+                    'description': 'Fetch one live Gitea issue by issue number for a tracked repository.',
+                    'arguments': {
+                        'project_id': 'optional tracked project id',
+                        'owner': 'optional repository owner override',
+                        'repo': 'optional repository name override',
+                        'issue_number': 'required integer issue number',
+                    },
+                }
+            )
+        if 'gitea_lookup_pull_request' in allowed:
+            specs.append(
+                {
+                    'name': 'gitea_lookup_pull_request',
+                    'description': 'Fetch one live Gitea pull request by PR number for a tracked repository.',
+                    'arguments': {
+                        'project_id': 'optional tracked project id',
+                        'owner': 'optional repository owner override',
+                        'repo': 'optional repository name override',
+                        'pr_number': 'required integer pull request number',
+                    },
+                }
+            )
+        return specs
+
+
+class LLMLiveToolExecutor:
+    """Resolve bounded live tool requests on behalf of the model."""
+
+    def __init__(self):
+        self.gitea_api = None
+        if settings.gitea_url and settings.gitea_token:
+            self.gitea_api = GiteaAPI(
+                token=settings.GITEA_TOKEN,
+                base_url=settings.GITEA_URL,
+                owner=settings.GITEA_OWNER,
+                repo=settings.GITEA_REPO or '',
+            )
+
+    async def execute(self, tool_name: str, arguments: dict, context: dict | None = None) -> dict:
+        """Execute one live tool request and normalize the result."""
+        if tool_name not in set(settings.llm_live_tool_allowlist):
+            return {'error': f'Tool {tool_name} is not enabled'}
+        if self.gitea_api is None:
+            return {'error': 'Gitea live tool execution is not configured'}
+        resolved = self._resolve_repository(arguments=arguments, context=context or {})
+        if resolved.get('error'):
+            return resolved
+        owner = resolved['owner']
+        repo = resolved['repo']
+
+        if tool_name == 'gitea_lookup_issue':
+            issue_number = arguments.get('issue_number')
+            if issue_number is None:
+                return {'error': 'issue_number is required'}
+            return await self.gitea_api.get_issue(issue_number=int(issue_number), owner=owner, repo=repo)
+
+        if tool_name == 'gitea_lookup_pull_request':
+            pr_number = arguments.get('pr_number')
+            if pr_number is None:
+                return {'error': 'pr_number is required'}
+            return await self.gitea_api.get_pull_request(pr_number=int(pr_number), owner=owner, repo=repo)
+
+        return {'error': f'Unsupported tool {tool_name}'}
+
+    def _resolve_repository(self, arguments: dict, context: dict) -> dict:
+        """Resolve repository owner/name from explicit args or tracked project context."""
+        owner = arguments.get('owner')
+        repo = arguments.get('repo')
+        if owner and repo:
+            return {'owner': owner, 'repo': repo}
+        project_id = arguments.get('project_id')
+        if project_id:
+            for project in context.get('projects', []):
+                if project.get('project_id') == project_id:
+                    repository = project.get('repository') or {}
+                    if repository.get('owner') and repository.get('name'):
+                        return {'owner': repository['owner'], 'repo': repository['name']}
+            state = context.get('repository') or {}
+            if context.get('project_id') == project_id and state.get('owner') and state.get('name'):
+                return {'owner': state['owner'], 'repo': state['name']}
+        repository = context.get('repository') or {}
+        if repository.get('owner') and repository.get('name'):
+            return {'owner': repository['owner'], 'repo': repository['name']}
+        return {'error': 'Could not resolve repository for tool request'}
+
+
+class LLMServiceClient:
+    """Call the configured LLM provider with consistent guardrails and tool payloads."""
+
+    def __init__(self, ollama_url: str | None = None, model: str | None = None):
+        self.ollama_url = (ollama_url or settings.ollama_url).rstrip('/')
+        self.model = model or settings.OLLAMA_MODEL
+        self.toolbox = LLMToolbox()
+        self.live_tool_executor = LLMLiveToolExecutor()
+
+    async def chat_with_trace(
+        self,
+        *,
+        stage: str,
+        system_prompt: str,
+        user_prompt: str,
+        tool_context_input: dict | None = None,
+        expect_json: bool = False,
+    ) -> tuple[str | None, dict]:
+        """Invoke the configured LLM and return both content and a structured trace."""
+        effective_system_prompt = self._compose_system_prompt(stage, system_prompt)
+        tool_context = self.toolbox.build_tool_context(stage=stage, context=tool_context_input)
+        live_tool_specs = self.toolbox.build_live_tool_specs(stage=stage, context=tool_context_input)
+        effective_user_prompt = self._compose_user_prompt(user_prompt, tool_context, live_tool_specs)
+        raw_responses: list[dict] = []
+        executed_tool_calls: list[dict] = []
+        current_user_prompt = effective_user_prompt
+        max_rounds = settings.llm_max_tool_call_rounds
+
+        for round_index in range(max_rounds + 1):
+            content, payload, error = await self._send_chat_request(
+                system_prompt=effective_system_prompt,
+                user_prompt=current_user_prompt,
+                expect_json=expect_json,
+            )
+            raw_responses.append(payload)
+            if content:
+                tool_request = self._extract_tool_request(content)
+                if tool_request and round_index < max_rounds:
+                    tool_name = tool_request.get('name')
+                    tool_arguments = tool_request.get('arguments') or {}
+                    tool_result = await self.live_tool_executor.execute(tool_name, tool_arguments, tool_context_input)
+                    executed_tool_calls.append(
+                        {
+                            'name': tool_name,
+                            'arguments': tool_arguments,
+                            'result': tool_result,
+                        }
+                    )
+                    current_user_prompt = self._compose_follow_up_prompt(user_prompt, tool_context, live_tool_specs, executed_tool_calls)
+                    continue
+                return content, {
+                    'stage': stage,
+                    'provider': 'ollama',
+                    'model': self.model,
+                    'system_prompt': effective_system_prompt,
+                    'user_prompt': current_user_prompt,
+                    'assistant_response': content,
+                    'raw_response': {
+                        'provider_response': raw_responses[-1],
+                        'provider_responses': raw_responses,
+                        'tool_context': tool_context,
+                        'live_tool_specs': live_tool_specs,
+                        'executed_tool_calls': executed_tool_calls,
+                    },
+                    'raw_responses': raw_responses,
+                    'fallback_used': False,
+                    'guardrails': self._guardrail_sections(stage),
+                    'tool_context': tool_context,
+                    'live_tool_specs': live_tool_specs,
+                    'executed_tool_calls': executed_tool_calls,
+                }
+            if error:
+                break
+
+        return None, {
+            'stage': stage,
+            'provider': 'ollama',
+            'model': self.model,
+            'system_prompt': effective_system_prompt,
+            'user_prompt': current_user_prompt,
+            'assistant_response': '',
+            'raw_response': {
+                'provider_response': raw_responses[-1] if raw_responses else {'error': 'No response'},
+                'provider_responses': raw_responses,
+                'tool_context': tool_context,
+                'live_tool_specs': live_tool_specs,
+                'executed_tool_calls': executed_tool_calls,
+            },
+            'raw_responses': raw_responses,
+            'fallback_used': True,
+            'guardrails': self._guardrail_sections(stage),
+            'tool_context': tool_context,
+            'live_tool_specs': live_tool_specs,
+            'executed_tool_calls': executed_tool_calls,
+        }
+
+    async def _send_chat_request(self, *, system_prompt: str, user_prompt: str, expect_json: bool) -> tuple[str | None, dict, str | None]:
+        """Send one outbound chat request to the configured model provider."""
+        request_payload = {
+            'model': self.model,
+            'stream': False,
+            'messages': [
+                {'role': 'system', 'content': system_prompt},
+                {'role': 'user', 'content': user_prompt},
+            ],
+        }
+        if expect_json:
+            request_payload['format'] = 'json'
+        try:
+            import aiohttp
+
+            async with aiohttp.ClientSession() as session:
+                async with session.post(f'{self.ollama_url}/api/chat', json=request_payload) as resp:
+                    payload = await resp.json()
+                    if 200 <= resp.status < 300:
+                        return (payload.get('message') or {}).get('content', ''), payload, None
+                    return None, payload, str(payload.get('error') or payload)
+        except Exception as exc:
+            return None, {'error': str(exc)}, str(exc)
+
+    def _compose_system_prompt(self, stage: str, stage_prompt: str) -> str:
+        """Merge the stage prompt with configured guardrails."""
+        sections = [stage_prompt.strip()] + self._guardrail_sections(stage)
+        return '\n\n'.join(section for section in sections if section)
+
+    def _guardrail_sections(self, stage: str) -> list[str]:
+        """Return all configured guardrail sections for one LLM stage."""
+        sections = []
+        if settings.llm_guardrail_prompt:
+            sections.append(f'Global guardrails:\n{settings.llm_guardrail_prompt}')
+        stage_specific = {
+            'request_interpretation': settings.llm_request_interpreter_guardrail_prompt,
+            'change_summary': settings.llm_change_summary_guardrail_prompt,
+            'project_naming': settings.llm_project_naming_guardrail_prompt,
+            'project_id_naming': settings.llm_project_id_guardrail_prompt,
+        }.get(stage)
+        if stage_specific:
+            sections.append(f'Stage-specific guardrails:\n{stage_specific}')
+        return sections
+
+    def _compose_user_prompt(self, prompt: str, tool_context: list[dict], live_tool_specs: list[dict] | None = None) -> str:
+        """Append tool payloads and live tool-call specs to the outbound user prompt."""
+        live_tool_specs = live_tool_specs if live_tool_specs is not None else []
+        sections = [prompt]
+        if not tool_context:
+            pass
+        else:
+            sections.append(
+                'Service-mediated tool outputs are available below. Treat them as authoritative read-only data supplied by the factory:\n'
+                f'{json.dumps(tool_context, indent=2, sort_keys=True)}'
+            )
+        if live_tool_specs:
+            sections.append(
+                'If you need additional live repository data, you may request exactly one tool call by responding with JSON shaped as '
+                '{"tool_request": {"name": "<tool name>", "arguments": {...}}}. '
+                'After tool results are returned, respond with the final answer instead of another tool request.\n'
+                f'Available live tools:\n{json.dumps(live_tool_specs, indent=2, sort_keys=True)}'
+            )
+        return '\n\n'.join(section for section in sections if section)
+
+    def _compose_follow_up_prompt(self, original_prompt: str, tool_context: list[dict], live_tool_specs: list[dict], executed_tool_calls: list[dict]) -> str:
+        """Build the follow-up user prompt after executing one or more live tool requests."""
+        sections = [self._compose_user_prompt(original_prompt, tool_context, live_tool_specs)]
+        sections.append(
+            'The service executed the requested live tool call(s). Use the tool result(s) below to produce the final answer. Do not request another tool call.\n'
+            f'{json.dumps(executed_tool_calls, indent=2, sort_keys=True)}'
+        )
+        return '\n\n'.join(sections)
+
+    def _extract_tool_request(self, content: str) -> dict | None:
+        """Return a normalized tool request when the model explicitly asks for one."""
+        try:
+            parsed = json.loads(content)
+        except Exception:
+            return None
+        if not isinstance(parsed, dict):
+            return None
+        tool_request = parsed.get('tool_request')
+        if not isinstance(tool_request, dict) or not tool_request.get('name'):
+            return None
+        return {
+            'name': str(tool_request.get('name')).strip(),
+            'arguments': tool_request.get('arguments') or {},
+        }
+
+    def get_runtime_configuration(self) -> dict:
+        """Return the active LLM runtime config, guardrails, and tool exposure."""
+        live_tool_stages = {
+            stage: settings.llm_live_tools_for_stage(stage)
+            for stage in self.toolbox.SUPPORTED_LIVE_TOOL_STAGES
+        }
+        return {
+            'provider': 'ollama',
+            'ollama_url': self.ollama_url,
+            'model': self.model,
+            'guardrails': {
+                'global': settings.llm_guardrail_prompt,
+                'request_interpretation': settings.llm_request_interpreter_guardrail_prompt,
+                'change_summary': settings.llm_change_summary_guardrail_prompt,
+                'project_naming': settings.llm_project_naming_guardrail_prompt,
+                'project_id_naming': settings.llm_project_id_guardrail_prompt,
+            },
+            'system_prompts': {
+                'project_naming': settings.llm_project_naming_system_prompt,
+                'project_id_naming': settings.llm_project_id_system_prompt,
+            },
+            'mediated_tools': settings.llm_tool_allowlist,
+            'live_tools': settings.llm_live_tool_allowlist,
+            'live_tool_stage_allowlist': settings.llm_live_tool_stage_allowlist,
+            'live_tool_stage_tool_map': settings.llm_live_tool_stage_tool_map,
+            'live_tools_by_stage': live_tool_stages,
+            'tool_context_limit': settings.llm_tool_context_limit,
+            'max_tool_call_rounds': settings.llm_max_tool_call_rounds,
+            'gitea_live_tools_configured': bool(settings.gitea_url and settings.gitea_token),
+        }

ai_software_factory/agents/orchestrator.py

@@ -3,9 +3,11 @@
 from __future__ import annotations
 
 import difflib
+import json
 import py_compile
 import re
 import subprocess
+from pathlib import PurePosixPath
 from typing import Optional
 from datetime import datetime
 
@@ -14,18 +16,27 @@ try:
     from .database_manager import DatabaseManager
     from .git_manager import GitManager
     from .gitea import GiteaAPI
+    from .llm_service import LLMServiceClient
     from .ui_manager import UIManager
 except ImportError:
     from config import settings
     from agents.database_manager import DatabaseManager
     from agents.git_manager import GitManager
     from agents.gitea import GiteaAPI
+    from agents.llm_service import LLMServiceClient
     from agents.ui_manager import UIManager
 
 
 class AgentOrchestrator:
     """Orchestrates the software generation process with full audit trail."""
 
+    REMOTE_READY_REPOSITORY_MODES = {'project', 'onboarded'}
+    REMOTE_READY_REPOSITORY_STATUSES = {'created', 'exists', 'ready', 'onboarded'}
+    GENERATED_TEXT_FILE_SUFFIXES = {'.py', '.md', '.txt', '.toml', '.yaml', '.yml', '.json', '.ini', '.cfg', '.sh', '.html', '.css', '.js', '.ts'}
+    GENERATED_TEXT_FILE_NAMES = {'README', 'README.md', '.gitignore', 'requirements.txt', 'pyproject.toml', 'Dockerfile', 'Containerfile', 'Makefile'}
+    MAX_WORKSPACE_CONTEXT_FILES = 20
+    MAX_WORKSPACE_CONTEXT_CHARS = 24000
+
     def __init__(
         self,
         project_id: str,
@@ -39,6 +50,7 @@ class AgentOrchestrator:
         existing_history=None,
         prompt_source_context: dict | None = None,
         prompt_routing: dict | None = None,
+        repo_name_override: str | None = None,
         related_issue_hint: dict | None = None,
     ):
         """Initialize orchestrator."""
@@ -58,8 +70,10 @@ class AgentOrchestrator:
         self.prompt_actor = prompt_actor
         self.prompt_source_context = prompt_source_context or {}
         self.prompt_routing = prompt_routing or {}
+        self.repo_name_override = repo_name_override
         self.existing_history = existing_history
         self.changed_files: list[str] = []
+        self.pending_code_changes: list[dict] = []
         self.gitea_api = GiteaAPI(
             token=settings.GITEA_TOKEN,
             base_url=settings.GITEA_URL,
@@ -68,12 +82,13 @@ class AgentOrchestrator:
         )
         self.project_root = settings.projects_root / project_id
         self.prompt_audit = None
-        self.repo_name = settings.gitea_repo or self.gitea_api.build_project_repo_name(project_id, project_name)
+        self.repo_name = settings.gitea_repo or self.gitea_api.build_project_repo_name(project_id, repo_name_override or project_name)
         self.repo_owner = settings.gitea_owner
         self.repo_url = None
         self.branch_name = self._build_pr_branch_name(project_id)
         self.active_pull_request = None
         self._gitea_username: str | None = None
+        existing_repository: dict | None = None
         hinted_issue_number = (related_issue_hint or {}).get('number') if related_issue_hint else None
         self.related_issue_number = hinted_issue_number if hinted_issue_number is not None else self._extract_issue_number(prompt_text)
         self.related_issue: dict | None = DatabaseManager._normalize_issue(related_issue_hint)
@@ -104,9 +119,12 @@ class AgentOrchestrator:
                 latest_ui = self.db_manager._get_latest_ui_snapshot_data(self.history.id)
                 repository = latest_ui.get('repository') if isinstance(latest_ui, dict) else None
                 if isinstance(repository, dict) and repository:
+                    existing_repository = dict(repository)
                     self.repo_owner = repository.get('owner') or self.repo_owner
                     self.repo_name = repository.get('name') or self.repo_name
                     self.repo_url = repository.get('url') or self.repo_url
+                    git_state = latest_ui.get('git') if isinstance(latest_ui.get('git'), dict) else {}
+                    self.branch_name = git_state.get('active_branch') or self.branch_name
             if self.prompt_text:
                 self.prompt_audit = self.db_manager.log_prompt_submission(
                     history_id=self.history.id,
@@ -123,18 +141,60 @@ class AgentOrchestrator:
         self.ui_manager.ui_data["project_root"] = str(self.project_root)
         self.ui_manager.ui_data["features"] = list(self.features)
         self.ui_manager.ui_data["tech_stack"] = list(self.tech_stack)
-        self.ui_manager.ui_data["repository"] = {
+        repository_ui = {
             "owner": self.repo_owner,
             "name": self.repo_name,
             "mode": "project" if settings.use_project_repositories else "shared",
             "status": "pending" if settings.use_project_repositories else "shared",
             "provider": "gitea",
         }
+        if existing_repository:
+            repository_ui.update(existing_repository)
+        self.ui_manager.ui_data["repository"] = repository_ui
         if self.related_issue:
             self.ui_manager.ui_data["related_issue"] = self.related_issue
         if self.active_pull_request:
             self.ui_manager.ui_data["pull_request"] = self.active_pull_request
 
+    def _repository_supports_remote_delivery(self, repository: dict | None = None) -> bool:
+        """Return whether repository metadata supports git push and PR delivery."""
+        repo = repository or self.ui_manager.ui_data.get('repository') or {}
+        return repo.get('mode') in self.REMOTE_READY_REPOSITORY_MODES and repo.get('status') in self.REMOTE_READY_REPOSITORY_STATUSES
+
+    def _static_files(self) -> dict[str, str]:
+        """Files that do not need prompt-specific generation."""
+        return {
+            ".gitignore": "__pycache__/\n*.pyc\n.venv/\n.pytest_cache/\n.mypy_cache/\n",
+        }
+
+    def _fallback_generated_files(self) -> dict[str, str]:
+        """Deterministic fallback files when LLM generation is unavailable."""
+        feature_section = "\n".join(f"- {feature}" for feature in self.features) or "- None specified"
+        tech_section = "\n".join(f"- {tech}" for tech in self.tech_stack) or "- Python"
+        return {
+            "README.md": (
+                f"# {self.project_name}\n\n"
+                f"{self.description}\n\n"
+                "## Features\n"
+                f"{feature_section}\n\n"
+                "## Tech Stack\n"
+                f"{tech_section}\n"
+            ),
+            "requirements.txt": "fastapi\nuvicorn\npytest\n",
+            "main.py": (
+                "from fastapi import FastAPI\n\n"
+                "app = FastAPI(title=\"Generated App\")\n\n"
+                "@app.get('/')\n"
+                "def read_root():\n"
+                f"    return {{'name': '{self.project_name}', 'status': 'generated', 'features': {self.features!r}}}\n"
+            ),
+            "tests/test_app.py": (
+                "from main import read_root\n\n"
+                "def test_read_root():\n"
+                f"    assert read_root()['name'] == '{self.project_name}'\n"
+            ),
+        }
+
     def _build_pr_branch_name(self, project_id: str) -> str:
         """Build a stable branch name used until the PR is merged."""
         return f"ai/{project_id}"
@@ -155,7 +215,7 @@ class AgentOrchestrator:
         """Persist the current generation plan as an inspectable trace."""
         if not self.db_manager or not self.history or not self.prompt_audit:
             return
-        planned_files = list(self._template_files().keys())
+        planned_files = list(self._static_files().keys()) + list(self._fallback_generated_files().keys())
         self.db_manager.log_llm_trace(
             project_id=self.project_id,
             history_id=self.history.id,
@@ -185,6 +245,175 @@ class AgentOrchestrator:
             fallback_used=False,
         )
 
+    def _is_safe_relative_path(self, path: str) -> bool:
+        """Return whether a generated file path is safe to write under the project root."""
+        normalized = str(PurePosixPath((path or '').strip()))
+        if not normalized or normalized in {'.', '..'}:
+            return False
+        if normalized.startswith('/') or normalized.startswith('../') or '/../' in normalized:
+            return False
+        if normalized.startswith('.git/'):
+            return False
+        return True
+
+    def _is_supported_generated_text_file(self, path: str) -> bool:
+        """Return whether the generated path is a supported text artifact."""
+        normalized = PurePosixPath(path)
+        if normalized.name in self.GENERATED_TEXT_FILE_NAMES:
+            return True
+        return normalized.suffix.lower() in self.GENERATED_TEXT_FILE_SUFFIXES
+
+    def _collect_workspace_context(self) -> dict:
+        """Collect a compact, text-only snapshot of the current project workspace."""
+        if not self.project_root.exists():
+            return {'has_existing_files': False, 'files': []}
+
+        files: list[dict] = []
+        total_chars = 0
+        for path in sorted(self.project_root.rglob('*')):
+            if not path.is_file():
+                continue
+            relative_path = path.relative_to(self.project_root).as_posix()
+            if relative_path == '.gitignore':
+                continue
+            if not self._is_safe_relative_path(relative_path) or not self._is_supported_generated_text_file(relative_path):
+                continue
+            try:
+                content = path.read_text(encoding='utf-8')
+            except (UnicodeDecodeError, OSError):
+                continue
+            remaining_chars = self.MAX_WORKSPACE_CONTEXT_CHARS - total_chars
+            if remaining_chars <= 0:
+                break
+            snippet = content[:remaining_chars]
+            files.append(
+                {
+                    'path': relative_path,
+                    'content': snippet,
+                    'truncated': len(snippet) < len(content),
+                }
+            )
+            total_chars += len(snippet)
+            if len(files) >= self.MAX_WORKSPACE_CONTEXT_FILES:
+                break
+        return {'has_existing_files': bool(files), 'files': files}
+
+    def _parse_generated_files(self, content: str | None) -> dict[str, str]:
+        """Parse an LLM file bundle response into relative-path/content pairs."""
+        if not content:
+            return {}
+        try:
+            parsed = json.loads(content)
+        except Exception:
+            return {}
+        generated: dict[str, str] = {}
+        if isinstance(parsed, dict) and isinstance(parsed.get('files'), list):
+            for item in parsed['files']:
+                if not isinstance(item, dict):
+                    continue
+                path = str(item.get('path') or '').strip()
+                file_content = item.get('content')
+                if (
+                    self._is_safe_relative_path(path)
+                    and self._is_supported_generated_text_file(path)
+                    and isinstance(file_content, str)
+                    and file_content.strip()
+                ):
+                    generated[path] = file_content.rstrip() + "\n"
+        elif isinstance(parsed, dict):
+            for path, file_content in parsed.items():
+                normalized_path = str(path).strip()
+                if (
+                    self._is_safe_relative_path(normalized_path)
+                    and self._is_supported_generated_text_file(normalized_path)
+                    and isinstance(file_content, str)
+                    and file_content.strip()
+                ):
+                    generated[normalized_path] = file_content.rstrip() + "\n"
+        return generated
+
+    async def _generate_prompt_driven_files(self) -> tuple[dict[str, str], dict | None, bool]:
+        """Use the configured LLM to generate prompt-specific project files."""
+        fallback_files = self._fallback_generated_files()
+        workspace_context = self._collect_workspace_context()
+        has_existing_files = bool(workspace_context.get('has_existing_files'))
+        if has_existing_files:
+            system_prompt = (
+                'You modify an existing software repository. '
+                'Return only JSON. Update the smallest necessary set of files to satisfy the new prompt. '
+                'Prefer editing existing files over inventing a new starter app. '
+                'Only return files that should be written. Omit unchanged files. '
+                'Use repository-relative paths and do not wrap the JSON in markdown fences.'
+            )
+            user_prompt = (
+                f"Project name: {self.project_name}\n"
+                f"Description: {self.description}\n"
+                f"Original prompt: {self.prompt_text or self.description}\n"
+                f"Requested features: {json.dumps(self.features)}\n"
+                f"Preferred tech stack: {json.dumps(self.tech_stack)}\n"
+                f"Related issue: {json.dumps(self.related_issue) if self.related_issue else 'null'}\n\n"
+                f"Current workspace snapshot:\n{json.dumps(workspace_context['files'], indent=2)}\n\n"
+                'Return JSON shaped as {"files": [{"path": "relative/path.py", "content": "..."}, ...]}. '
+                'Each file path must be relative to the repository root.'
+            )
+        else:
+            system_prompt = (
+                'You generate small but concrete starter projects. '
+                'Return only JSON. Provide production-like but compact code that directly reflects the user request. '
+                'Include the files README.md, requirements.txt, main.py, and tests/test_app.py. '
+                'Use FastAPI for Python web requests unless the prompt clearly demands something else. '
+                'The test must verify a real behavior from main.py. '
+                'Do not wrap the JSON in markdown fences.'
+            )
+            user_prompt = (
+                f"Project name: {self.project_name}\n"
+                f"Description: {self.description}\n"
+                f"Original prompt: {self.prompt_text or self.description}\n"
+                f"Requested features: {json.dumps(self.features)}\n"
+                f"Preferred tech stack: {json.dumps(self.tech_stack)}\n"
+                f"Related issue: {json.dumps(self.related_issue) if self.related_issue else 'null'}\n\n"
+                'Return JSON shaped as {"files": [{"path": "README.md", "content": "..."}, ...]}. '
+                'At minimum include README.md, requirements.txt, main.py, and tests/test_app.py.'
+            )
+        content, trace = await LLMServiceClient().chat_with_trace(
+            stage='generation_plan',
+            system_prompt=system_prompt,
+            user_prompt=user_prompt,
+            tool_context_input={
+                'project_id': self.project_id,
+                'project_name': self.project_name,
+                'repository': self.ui_manager.ui_data.get('repository'),
+                'related_issue': self.related_issue,
+                'workspace_files': workspace_context.get('files', []),
+            },
+            expect_json=True,
+        )
+        raw_generated_paths = self._extract_raw_generated_paths(content)
+        generated_files = self._parse_generated_files(content)
+        accepted_paths = list(generated_files.keys())
+        rejected_paths = [path for path in raw_generated_paths if path not in accepted_paths]
+        generation_debug = {
+            'raw_paths': raw_generated_paths,
+            'accepted_paths': accepted_paths,
+            'rejected_paths': rejected_paths,
+            'existing_workspace': has_existing_files,
+        }
+        self.ui_manager.ui_data['generation_debug'] = generation_debug
+        self._append_log(
+            'LLM returned file candidates: '
+            f"raw={raw_generated_paths or []}; accepted={accepted_paths or []}; rejected={rejected_paths or []}."
+        )
+        self._log_system_debug(
+            'generation',
+            'LLM file candidates '
+            f"raw={raw_generated_paths or []}; accepted={accepted_paths or []}; rejected={rejected_paths or []}; "
+            f"existing_workspace={has_existing_files}",
+        )
+        if has_existing_files:
+            return generated_files, trace, True
+        merged_files = {**fallback_files, **generated_files}
+        return merged_files, trace, False
+
     async def _sync_issue_context(self) -> None:
         """Sync repository issues and resolve a linked issue from the prompt when present."""
         if not self.db_manager or not self.history:
@@ -209,6 +438,14 @@ class AgentOrchestrator:
             self.db_manager.attach_issue_to_prompt(self.prompt_audit.id, self.related_issue)
 
     async def _ensure_remote_repository(self) -> None:
+        repository = self.ui_manager.ui_data.get("repository") or {}
+        if self._repository_supports_remote_delivery(repository):
+            repository.setdefault("provider", "gitea")
+            repository.setdefault("status", "ready")
+            if repository.get("url"):
+                self.repo_url = repository.get("url")
+            self.ui_manager.ui_data["repository"] = repository
+            return
         if not settings.use_project_repositories:
             self.ui_manager.ui_data["repository"]["status"] = "shared"
             if settings.gitea_repo:
@@ -300,9 +537,7 @@ class AgentOrchestrator:
     async def _push_branch(self, branch: str) -> dict | None:
         """Push a branch to the configured project repository when available."""
         repository = self.ui_manager.ui_data.get('repository') or {}
-        if repository.get('mode') != 'project':
-            return None
-        if repository.get('status') not in {'created', 'exists', 'ready'}:
+        if not self._repository_supports_remote_delivery(repository):
             return None
         if not settings.gitea_token or not self.repo_owner or not self.repo_name:
             return None
@@ -349,7 +584,7 @@ class AgentOrchestrator:
             self.ui_manager.ui_data['pull_request'] = self.active_pull_request
             return self.active_pull_request
         repository = self.ui_manager.ui_data.get('repository') or {}
-        if repository.get('mode') != 'project' or repository.get('status') not in {'created', 'exists', 'ready'}:
+        if not self._repository_supports_remote_delivery(repository):
             return None
 
         title = f"AI delivery for {self.project_name}"
@@ -390,9 +625,7 @@ class AgentOrchestrator:
     async def _push_remote_commit(self, commit_hash: str, commit_message: str, changed_files: list[str], base_commit: str | None) -> dict | None:
         """Push the local commit to the provisioned Gitea repository and build browser links."""
         repository = self.ui_manager.ui_data.get("repository") or {}
-        if repository.get("mode") != "project":
-            return None
-        if repository.get("status") not in {"created", "exists", "ready"}:
+        if not self._repository_supports_remote_delivery(repository):
             return None
         push_result = await self._push_branch(self.branch_name)
         if push_result is None:
@@ -434,6 +667,35 @@ class AgentOrchestrator:
         if self.db_manager and self.history:
             self.db_manager._log_action(self.history.id, "INFO", message)
 
+    def _log_system_debug(self, component: str, message: str, level: str = 'INFO') -> None:
+        """Persist a system-level debug breadcrumb for generation and git decisions."""
+        if not self.db_manager:
+            return
+        self.db_manager.log_system_event(component=component, level=level, message=f"{self.project_id}: {message}")
+
+    def _extract_raw_generated_paths(self, content: str | None) -> list[str]:
+        """Return all file paths proposed by the LLM response before safety filtering."""
+        if not content:
+            return []
+        try:
+            parsed = json.loads(content)
+        except Exception:
+            return []
+        raw_paths: list[str] = []
+        if isinstance(parsed, dict) and isinstance(parsed.get('files'), list):
+            for item in parsed['files']:
+                if not isinstance(item, dict):
+                    continue
+                path = str(item.get('path') or '').strip()
+                if path:
+                    raw_paths.append(path)
+        elif isinstance(parsed, dict):
+            for path in parsed.keys():
+                normalized_path = str(path).strip()
+                if normalized_path:
+                    raw_paths.append(normalized_path)
+        return raw_paths
+
     def _update_progress(self, progress: int, step: str, message: str) -> None:
         self.progress = progress
         self.current_step = step
@@ -452,50 +714,20 @@ class AgentOrchestrator:
         target.parent.mkdir(parents=True, exist_ok=True)
         change_type = "UPDATE" if target.exists() else "CREATE"
         previous_content = target.read_text(encoding="utf-8") if target.exists() else ""
+        if previous_content == content:
+            return
         diff_text = self._build_diff_text(relative_path, previous_content, content)
         target.write_text(content, encoding="utf-8")
         self.changed_files.append(relative_path)
-        if self.db_manager and self.history:
-            self.db_manager.log_code_change(
-                project_id=self.project_id,
-                change_type=change_type,
-                file_path=relative_path,
-                actor="orchestrator",
-                actor_type="agent",
-                details=f"{change_type.title()}d generated artifact {relative_path}",
-                history_id=self.history.id,
-                prompt_id=self.prompt_audit.id if self.prompt_audit else None,
-                diff_summary=f"Wrote {len(content.splitlines())} lines to {relative_path}",
-                diff_text=diff_text,
-            )
-
-    def _template_files(self) -> dict[str, str]:
-        feature_section = "\n".join(f"- {feature}" for feature in self.features) or "- None specified"
-        tech_section = "\n".join(f"- {tech}" for tech in self.tech_stack) or "- Python"
-        return {
-            ".gitignore": "__pycache__/\n*.pyc\n.venv/\n.pytest_cache/\n.mypy_cache/\n",
-            "README.md": (
-                f"# {self.project_name}\n\n"
-                f"{self.description}\n\n"
-                "## Features\n"
-                f"{feature_section}\n\n"
-                "## Tech Stack\n"
-                f"{tech_section}\n"
-            ),
-            "requirements.txt": "fastapi\nuvicorn\npytest\n",
-            "main.py": (
-                "from fastapi import FastAPI\n\n"
-                "app = FastAPI(title=\"Generated App\")\n\n"
-                "@app.get('/')\n"
-                "def read_root():\n"
-                f"    return {{'name': '{self.project_name}', 'status': 'generated', 'features': {self.features!r}}}\n"
-            ),
-            "tests/test_app.py": (
-                "from main import read_root\n\n"
-                "def test_read_root():\n"
-                f"    assert read_root()['name'] == '{self.project_name}'\n"
-            ),
+        self.pending_code_changes.append(
+            {
+                'change_type': change_type,
+                'file_path': relative_path,
+                'details': f"{change_type.title()}d generated artifact {relative_path}",
+                'diff_summary': f"Wrote {len(content.splitlines())} lines to {relative_path}",
+                'diff_text': diff_text,
             }
+        )
 
     async def run(self) -> dict:
         """Run the software generation process with full audit logging."""
@@ -586,18 +818,37 @@ class AgentOrchestrator:
     async def _create_project_structure(self) -> None:
         """Create initial project structure."""
         self.project_root.mkdir(parents=True, exist_ok=True)
-        for relative_path, content in self._template_files().items():
-            if relative_path.startswith("main.py") or relative_path.startswith("tests/"):
-                continue
+        for relative_path, content in self._static_files().items():
             self._write_file(relative_path, content)
         self._append_log(f"Project structure created under {self.project_root}.")
 
     async def _generate_code(self) -> None:
         """Generate code using Ollama."""
-        for relative_path, content in self._template_files().items():
-            if relative_path in {"main.py", "tests/test_app.py"}:
+        change_count_before = len(self.pending_code_changes)
+        generated_files, trace, editing_existing_workspace = await self._generate_prompt_driven_files()
+        for relative_path, content in generated_files.items():
             self._write_file(relative_path, content)
-        self._append_log("Application entrypoint and smoke test generated.")
+        if editing_existing_workspace and len(self.pending_code_changes) == change_count_before:
+            raise RuntimeError('The LLM response did not produce any file changes for the existing project.')
+        fallback_used = bool(trace and trace.get('fallback_used')) or trace is None
+        if self.db_manager and self.history and self.prompt_audit and trace:
+            self.db_manager.log_llm_trace(
+                project_id=self.project_id,
+                history_id=self.history.id,
+                prompt_id=self.prompt_audit.id,
+                stage='code_generation',
+                provider=trace.get('provider', 'ollama'),
+                model=trace.get('model', settings.OLLAMA_MODEL),
+                system_prompt=trace.get('system_prompt', ''),
+                user_prompt=trace.get('user_prompt', self.prompt_text or self.description),
+                assistant_response=trace.get('assistant_response', ''),
+                raw_response=trace.get('raw_response'),
+                fallback_used=fallback_used,
+            )
+        if fallback_used:
+            self._append_log('LLM code generation was unavailable; used deterministic scaffolding fallback.')
+        else:
+            self._append_log('Application files generated from the prompt with the configured LLM.')
 
     async def _run_tests(self) -> None:
         """Run tests for the generated code."""
@@ -608,11 +859,25 @@ class AgentOrchestrator:
     async def _commit_to_git(self) -> None:
         """Commit changes to git."""
         unique_files = list(dict.fromkeys(self.changed_files))
+        git_debug = self.ui_manager.ui_data.setdefault('git', {})
         if not unique_files:
+            git_debug.update({
+                'commit_status': 'skipped',
+                'early_exit_reason': 'changed_files_empty',
+                'candidate_files': [],
+            })
+            self._append_log('Git commit skipped: no generated files were marked as changed.')
+            self._log_system_debug('git', 'Commit exited early because changed_files was empty.')
             return
         if not self.git_manager.is_git_available():
-            self.ui_manager.ui_data.setdefault('git', {})['error'] = 'git executable is not available in PATH'
+            git_debug.update({
+                'commit_status': 'error',
+                'early_exit_reason': 'git_unavailable',
+                'candidate_files': unique_files,
+                'error': 'git executable is not available in PATH',
+            })
             self._append_log('Git commit skipped: git executable is not available in PATH')
+            self._log_system_debug('git', 'Commit exited early because git is unavailable.', level='ERROR')
             return
 
         try:
@@ -620,7 +885,23 @@ class AgentOrchestrator:
                 self.git_manager.init_repo()
             base_commit = self.git_manager.current_head_or_none()
             self.git_manager.add_files(unique_files)
-            if not self.git_manager.get_status():
+            status_after_add = self.git_manager.get_status()
+            if not status_after_add:
+                git_debug.update({
+                    'commit_status': 'skipped',
+                    'early_exit_reason': 'clean_after_staging',
+                    'candidate_files': unique_files,
+                    'status_after_add': '',
+                })
+                self._append_log(
+                    'Git commit skipped: working tree was clean after staging candidate files '
+                    f'{unique_files}. No repository diff was created.'
+                )
+                self._log_system_debug(
+                    'git',
+                    'Commit exited early because git status was clean after staging '
+                    f'files={unique_files}',
+                )
                 return
 
             commit_message = f"AI generation for prompt: {self.project_name}"
@@ -633,11 +914,17 @@ class AgentOrchestrator:
                 "scope": "local",
                 "branch": self.branch_name,
             }
+            git_debug.update({
+                'commit_status': 'committed',
+                'early_exit_reason': None,
+                'candidate_files': unique_files,
+                'status_after_add': status_after_add,
+            })
             remote_record = None
             try:
                 remote_record = await self._push_remote_commit(commit_hash, commit_message, unique_files, base_commit)
             except (RuntimeError, subprocess.CalledProcessError, FileNotFoundError) as remote_exc:
-                self.ui_manager.ui_data.setdefault("git", {})["remote_error"] = str(remote_exc)
+                git_debug["remote_error"] = str(remote_exc)
                 self._append_log(f"Remote git push skipped: {remote_exc}")
 
             if remote_record:
@@ -647,8 +934,8 @@ class AgentOrchestrator:
                 if remote_record.get('pull_request'):
                     commit_record['pull_request'] = remote_record['pull_request']
                     self.ui_manager.ui_data['pull_request'] = remote_record['pull_request']
-            self.ui_manager.ui_data.setdefault("git", {})["latest_commit"] = commit_record
-            self.ui_manager.ui_data.setdefault("git", {})["commits"] = [commit_record]
+            git_debug["latest_commit"] = commit_record
+            git_debug["commits"] = [commit_record]
             self._append_log(f"Recorded git commit {commit_hash[:12]} for generated files.")
             if self.db_manager:
                 self.db_manager.log_commit(
@@ -666,6 +953,23 @@ class AgentOrchestrator:
                     remote_status=remote_record.get("status") if remote_record else "local-only",
                     related_issue=self.related_issue,
                 )
+                for change in self.pending_code_changes:
+                    self.db_manager.log_code_change(
+                        project_id=self.project_id,
+                        change_type=change['change_type'],
+                        file_path=change['file_path'],
+                        actor='orchestrator',
+                        actor_type='agent',
+                        details=change['details'],
+                        history_id=self.history.id if self.history else None,
+                        prompt_id=self.prompt_audit.id if self.prompt_audit else None,
+                        diff_summary=change.get('diff_summary'),
+                        diff_text=change.get('diff_text'),
+                        commit_hash=commit_hash,
+                        remote_status=remote_record.get('status') if remote_record else 'local-only',
+                        branch=self.branch_name,
+                    )
+                self.pending_code_changes.clear()
                 if self.related_issue:
                     self.db_manager.log_issue_work(
                         project_id=self.project_id,
@@ -677,7 +981,12 @@ class AgentOrchestrator:
                         commit_url=remote_record.get('commit_url') if remote_record else None,
                     )
         except (RuntimeError, subprocess.CalledProcessError, FileNotFoundError) as exc:
-            self.ui_manager.ui_data.setdefault("git", {})["error"] = str(exc)
+            git_debug.update({
+                'commit_status': 'error',
+                'early_exit_reason': 'commit_exception',
+                'candidate_files': unique_files,
+                'error': str(exc),
+            })
             self._append_log(f"Git commit skipped: {exc}")
 
     async def _create_pr(self) -> None:

ai_software_factory/agents/request_interpreter.py

@@ -7,16 +7,43 @@ import re
 
 try:
     from ..config import settings
+    from .gitea import GiteaAPI
+    from .llm_service import LLMServiceClient
 except ImportError:
     from config import settings
+    from agents.gitea import GiteaAPI
+    from agents.llm_service import LLMServiceClient
 
 
 class RequestInterpreter:
     """Use Ollama to turn free-form text into a structured software request."""
 
+    REQUEST_PREFIX_WORDS = {
+        'a', 'an', 'app', 'application', 'build', 'create', 'dashboard', 'develop', 'design', 'for', 'generate',
+        'internal', 'make', 'me', 'modern', 'need', 'new', 'our', 'platform', 'please', 'project', 'service',
+        'simple', 'site', 'start', 'system', 'the', 'tool', 'us', 'want', 'web', 'website', 'with',
+    }
+
+    REPO_NOISE_WORDS = REQUEST_PREFIX_WORDS | {'and', 'from', 'into', 'on', 'that', 'this', 'to'}
+    GENERIC_PROJECT_NAME_WORDS = {
+        'app', 'application', 'harness', 'platform', 'project', 'purpose', 'service', 'solution', 'suite', 'system', 'test', 'tool',
+    }
+    PLACEHOLDER_PROJECT_NAME_WORDS = {
+        'generated project', 'new project', 'project', 'temporary name', 'temp name', 'placeholder', 'untitled project',
+    }
+
     def __init__(self, ollama_url: str | None = None, model: str | None = None):
         self.ollama_url = (ollama_url or settings.ollama_url).rstrip('/')
         self.model = model or settings.OLLAMA_MODEL
+        self.llm_client = LLMServiceClient(ollama_url=self.ollama_url, model=self.model)
+        self.gitea_api = None
+        if settings.gitea_url and settings.gitea_token:
+            self.gitea_api = GiteaAPI(
+                token=settings.GITEA_TOKEN,
+                base_url=settings.GITEA_URL,
+                owner=settings.GITEA_OWNER,
+                repo=settings.GITEA_REPO or '',
+            )
 
     async def interpret(self, prompt_text: str, context: dict | None = None) -> dict:
         """Interpret free-form text into the request shape expected by the orchestrator."""
@@ -49,48 +76,48 @@ class RequestInterpreter:
                 f"User prompt:\n{normalized}"
             )
 
-        try:
-            import aiohttp
-
-            async with aiohttp.ClientSession() as session:
-                async with session.post(
-                    f'{self.ollama_url}/api/chat',
-                    json={
-                        'model': self.model,
-                        'stream': False,
-                        'format': 'json',
-                        'messages': [
-                            {
-                                'role': 'system',
-                                'content': system_prompt,
-                            },
-                            {'role': 'user', 'content': user_prompt},
+        content, trace = await self.llm_client.chat_with_trace(
+            stage='request_interpretation',
+            system_prompt=system_prompt,
+            user_prompt=user_prompt,
+            tool_context_input={
+                'projects': compact_context.get('projects', []),
+                'open_issues': [
+                    issue
+                    for project in compact_context.get('projects', [])
+                    for issue in project.get('open_issues', [])
                 ],
+                'recent_chat_history': compact_context.get('recent_chat_history', []),
             },
-                ) as resp:
-                    payload = await resp.json()
-                    if 200 <= resp.status < 300:
-                        content = payload.get('message', {}).get('content', '')
+            expect_json=True,
+        )
         if content:
+            try:
                 parsed = json.loads(content)
                 interpreted = self._normalize_interpreted_request(parsed, normalized)
                 routing = self._normalize_routing(parsed.get('routing'), interpreted, compact_context)
-                            return interpreted, {
-                                'stage': 'request_interpretation',
-                                'provider': 'ollama',
-                                'model': self.model,
-                                'system_prompt': system_prompt,
-                                'user_prompt': user_prompt,
-                                'assistant_response': content,
-                                'raw_response': payload,
-                                'routing': routing,
-                                'context_excerpt': compact_context,
-                                'fallback_used': False,
-                            }
+                if routing.get('intent') == 'continue_project' and routing.get('project_name'):
+                    interpreted['name'] = routing['project_name']
+                naming_trace = None
+                if routing.get('intent') == 'new_project':
+                    interpreted, routing, naming_trace = await self._refine_new_project_identity(
+                        prompt_text=normalized,
+                        interpreted=interpreted,
+                        routing=routing,
+                        context=compact_context,
+                    )
+                trace['routing'] = routing
+                trace['context_excerpt'] = compact_context
+                if naming_trace is not None:
+                    trace['project_naming'] = naming_trace
+                return interpreted, trace
             except Exception:
                 pass
 
         interpreted, routing = self._heuristic_fallback(normalized, compact_context)
+        if routing.get('intent') == 'new_project':
+            constraints = await self._collect_project_identity_constraints(compact_context)
+            routing['repo_name'] = self._ensure_unique_repo_name(routing.get('repo_name') or interpreted.get('name') or 'project', constraints['repo_names'])
         return interpreted, {
             'stage': 'request_interpretation',
             'provider': 'heuristic',
@@ -98,12 +125,88 @@ class RequestInterpreter:
             'system_prompt': system_prompt,
             'user_prompt': user_prompt,
             'assistant_response': json.dumps({'request': interpreted, 'routing': routing}),
-            'raw_response': {'fallback': 'heuristic'},
+            'raw_response': {'fallback': 'heuristic', 'llm_trace': trace.get('raw_response') if isinstance(trace, dict) else None},
             'routing': routing,
             'context_excerpt': compact_context,
+            'guardrails': trace.get('guardrails') if isinstance(trace, dict) else [],
+            'tool_context': trace.get('tool_context') if isinstance(trace, dict) else [],
             'fallback_used': True,
         }
 
+    async def _refine_new_project_identity(
+        self,
+        *,
+        prompt_text: str,
+        interpreted: dict,
+        routing: dict,
+        context: dict,
+    ) -> tuple[dict, dict, dict | None]:
+        """Refine project and repository naming for genuinely new work."""
+        constraints = await self._collect_project_identity_constraints(context)
+        user_prompt = (
+            f"Original user prompt:\n{prompt_text}\n\n"
+            f"Draft structured request:\n{json.dumps(interpreted, indent=2)}\n\n"
+            f"Tracked project names to avoid reusing unless the user clearly wants them:\n{json.dumps(sorted(constraints['project_names']))}\n\n"
+            f"Repository slugs already reserved in tracked projects or Gitea:\n{json.dumps(sorted(constraints['repo_names']))}\n\n"
+            "Suggest the best project display name and repository slug for this new project."
+        )
+        content, trace = await self.llm_client.chat_with_trace(
+            stage='project_naming',
+            system_prompt=settings.llm_project_naming_system_prompt,
+            user_prompt=user_prompt,
+            tool_context_input={
+                'projects': context.get('projects', []),
+            },
+            expect_json=True,
+        )
+        if content:
+            try:
+                fallback_name = self._preferred_project_name_fallback(prompt_text, interpreted.get('name'))
+                parsed = json.loads(content)
+                project_name, repo_name = self._normalize_project_identity(
+                    parsed,
+                    fallback_name=fallback_name,
+                )
+                repo_name = self._ensure_unique_repo_name(repo_name, constraints['repo_names'])
+                interpreted['name'] = project_name
+                routing['project_name'] = project_name
+                routing['repo_name'] = repo_name
+                return interpreted, routing, trace
+            except Exception:
+                pass
+
+        fallback_name = self._preferred_project_name_fallback(prompt_text, interpreted.get('name'))
+        routing['project_name'] = fallback_name
+        routing['repo_name'] = self._ensure_unique_repo_name(self._derive_repo_name(fallback_name), constraints['repo_names'])
+        return interpreted, routing, trace
+
+    async def _collect_project_identity_constraints(self, context: dict) -> dict[str, set[str]]:
+        """Collect reserved project names and repository slugs from tracked state and Gitea."""
+        project_names: set[str] = set()
+        repo_names: set[str] = set()
+        for project in context.get('projects', []):
+            if project.get('name'):
+                project_names.add(str(project.get('name')).strip())
+            repository = project.get('repository') or {}
+            if repository.get('name'):
+                repo_names.add(str(repository.get('name')).strip())
+        repo_names.update(await self._load_remote_repo_names())
+        return {
+            'project_names': project_names,
+            'repo_names': repo_names,
+        }
+
+    async def _load_remote_repo_names(self) -> set[str]:
+        """Load current Gitea repository names when live credentials are available."""
+        if settings.gitea_repo:
+            return {settings.gitea_repo}
+        if self.gitea_api is None or not settings.gitea_owner:
+            return set()
+        repos = await self.gitea_api.list_repositories(owner=settings.gitea_owner)
+        if not isinstance(repos, list):
+            return set()
+        return {str(repo.get('name')).strip() for repo in repos if repo.get('name')}
+
     def _normalize_interpreted_request(self, interpreted: dict, original_prompt: str) -> dict:
         """Normalize LLM output into the required request shape."""
         request_payload = interpreted.get('request') if isinstance(interpreted.get('request'), dict) else interpreted
@@ -164,14 +267,26 @@ class RequestInterpreter:
                 matched_project = project
                 break
         intent = str(routing.get('intent') or '').strip() or ('continue_project' if matched_project else 'new_project')
-        return {
+        if matched_project is None and intent == 'continue_project':
+            recent_chat_history = context.get('recent_chat_history', [])
+            recent_project_id = recent_chat_history[0].get('project_id') if recent_chat_history else None
+            if recent_project_id:
+                matched_project = next(
+                    (project for project in context.get('projects', []) if project.get('project_id') == recent_project_id),
+                    None,
+                )
+        normalized = {
             'intent': intent,
             'project_id': matched_project.get('project_id') if matched_project else project_id,
             'project_name': matched_project.get('name') if matched_project else (project_name or interpreted.get('name')),
+            'repo_name': routing.get('repo_name') if intent == 'new_project' else None,
             'issue_number': issue_number,
             'confidence': routing.get('confidence') or ('medium' if matched_project else 'low'),
             'reasoning_summary': routing.get('reasoning_summary') or ('Matched prior project context' if matched_project else 'No strong prior project match found'),
         }
+        if normalized['intent'] == 'new_project' and not normalized['repo_name']:
+            normalized['repo_name'] = self._derive_repo_name(normalized['project_name'] or interpreted.get('name') or 'Generated Project')
+        return normalized
 
     def _normalize_list(self, value) -> list[str]:
         if isinstance(value, list):
@@ -190,13 +305,22 @@ class RequestInterpreter:
         noun_phrase = re.search(
             r'(?:build|create|start|make|develop|generate|design|need|want)\s+'
             r'(?:me\s+|us\s+|an?\s+|the\s+|new\s+|internal\s+|simple\s+|lightweight\s+|modern\s+|web\s+|mobile\s+)*'
-            r'([a-z0-9][a-z0-9\s-]{2,80}?(?:portal|dashboard|app|application|service|tool|system|platform|api|bot|assistant|website|site|workspace|tracker|manager))\b',
+            r'([a-z0-9][a-z0-9\s-]{2,80}?(?:portal|dashboard|app|application|service|tool|system|platform|api|bot|assistant|website|site|workspace|tracker|manager|harness|runner|framework|suite|pipeline|lab))\b',
             first_line,
             flags=re.IGNORECASE,
         )
         if noun_phrase:
             return self._humanize_name(noun_phrase.group(1))
 
+        focused_phrase = re.search(
+            r'(?:purpose\s+is\s+to\s+create\s+(?:an?\s+)?)'
+            r'([a-z0-9][a-z0-9\s-]{2,80}?(?:portal|dashboard|app|application|service|tool|system|platform|api|bot|assistant|website|site|workspace|tracker|manager|harness|runner|framework|suite|pipeline|lab))\b',
+            first_line,
+            flags=re.IGNORECASE,
+        )
+        if focused_phrase:
+            return self._humanize_name(focused_phrase.group(1))
+
         cleaned = re.sub(r'[^A-Za-z0-9 ]+', ' ', first_line)
         stopwords = {
             'build', 'create', 'start', 'make', 'develop', 'generate', 'design', 'need', 'want', 'please', 'for', 'our', 'with', 'that', 'this',
@@ -211,6 +335,7 @@ class RequestInterpreter:
         """Normalize a candidate project name into a readable title."""
         cleaned = re.sub(r'[^A-Za-z0-9\s-]+', ' ', raw_name).strip(' -')
         cleaned = re.sub(r'\s+', ' ', cleaned)
+        cleaned = self._trim_request_prefix(cleaned)
         special_upper = {'api', 'crm', 'erp', 'cms', 'hr', 'it', 'ui', 'qa'}
         words = []
         for word in cleaned.split()[:6]:
@@ -218,6 +343,102 @@ class RequestInterpreter:
             words.append(lowered.upper() if lowered in special_upper else lowered.capitalize())
         return ' '.join(words) or 'Generated Project'
 
+    def _trim_request_prefix(self, candidate: str) -> str:
+        """Remove leading request phrasing from model-produced names and slugs."""
+        tokens = [token for token in re.split(r'[-\s]+', candidate or '') if token]
+        while tokens and tokens[0].lower() in self.REQUEST_PREFIX_WORDS:
+            tokens.pop(0)
+        trimmed = ' '.join(tokens).strip()
+        return trimmed or candidate.strip()
+
+    def _derive_repo_name(self, project_name: str) -> str:
+        """Derive a repository slug from a human-readable project name."""
+        preferred_name = self._trim_request_prefix((project_name or 'project').strip())
+        preferred = preferred_name.lower().replace(' ', '-')
+        sanitized = ''.join(ch if ch.isalnum() or ch in {'-', '_'} else '-' for ch in preferred)
+        while '--' in sanitized:
+            sanitized = sanitized.replace('--', '-')
+        return sanitized.strip('-') or 'project'
+
+    def _should_use_repo_name_candidate(self, candidate: str, project_name: str) -> bool:
+        """Return whether a model-proposed repo slug is concise enough to trust directly."""
+        cleaned = self._trim_request_prefix(re.sub(r'[^A-Za-z0-9\s_-]+', ' ', candidate or '').strip())
+        if not cleaned:
+            return False
+        candidate_tokens = [token.lower() for token in re.split(r'[-\s_]+', cleaned) if token]
+        if not candidate_tokens:
+            return False
+        if len(candidate_tokens) > 6:
+            return False
+        noise_count = sum(1 for token in candidate_tokens if token in self.REPO_NOISE_WORDS)
+        if noise_count >= 2:
+            return False
+        if len('-'.join(candidate_tokens)) > 40:
+            return False
+        project_tokens = {
+            token.lower()
+            for token in re.split(r'[-\s_]+', project_name or '')
+            if token and token.lower() not in self.REPO_NOISE_WORDS
+        }
+        if project_tokens:
+            overlap = sum(1 for token in candidate_tokens if token in project_tokens)
+            if overlap == 0:
+                return False
+        return True
+
+    def _should_use_project_name_candidate(self, candidate: str, fallback_name: str) -> bool:
+        """Return whether a model-proposed project title is concrete enough to trust."""
+        cleaned = self._trim_request_prefix(re.sub(r'[^A-Za-z0-9\s-]+', ' ', candidate or '').strip())
+        if not cleaned:
+            return False
+        candidate_tokens = [token.lower() for token in re.split(r'[-\s]+', cleaned) if token]
+        if not candidate_tokens:
+            return False
+        if len(candidate_tokens) == 1 and candidate_tokens[0] in self.GENERIC_PROJECT_NAME_WORDS:
+            return False
+        if all(token in self.GENERIC_PROJECT_NAME_WORDS for token in candidate_tokens):
+            return False
+        fallback_tokens = {
+            token.lower() for token in re.split(r'[-\s]+', fallback_name or '') if token and token.lower() not in self.REPO_NOISE_WORDS
+        }
+        if fallback_tokens and len(candidate_tokens) <= 2:
+            overlap = sum(1 for token in candidate_tokens if token in fallback_tokens)
+            if overlap == 0 and any(token in self.GENERIC_PROJECT_NAME_WORDS for token in candidate_tokens):
+                return False
+        return True
+
+    def _preferred_project_name_fallback(self, prompt_text: str, interpreted_name: str | None) -> str:
+        """Pick the best fallback title when the earlier interpretation produced a placeholder."""
+        interpreted_clean = self._humanize_name(str(interpreted_name or '').strip()) if interpreted_name else ''
+        normalized_interpreted = interpreted_clean.lower()
+        if normalized_interpreted and normalized_interpreted not in self.PLACEHOLDER_PROJECT_NAME_WORDS:
+            if not (len(normalized_interpreted.split()) == 1 and normalized_interpreted in self.GENERIC_PROJECT_NAME_WORDS):
+                return interpreted_clean
+        return self._derive_name(prompt_text)
+
+    def _ensure_unique_repo_name(self, repo_name: str, reserved_names: set[str]) -> str:
+        """Choose a repository slug that does not collide with tracked or remote repositories."""
+        base_name = self._derive_repo_name(repo_name)
+        if base_name not in reserved_names:
+            return base_name
+        suffix = 2
+        while f'{base_name}-{suffix}' in reserved_names:
+            suffix += 1
+        return f'{base_name}-{suffix}'
+
+    def _normalize_project_identity(self, payload: dict, fallback_name: str) -> tuple[str, str]:
+        """Normalize model-proposed project and repository naming."""
+        fallback_project_name = self._humanize_name(str(fallback_name or 'Generated Project'))
+        project_candidate = str(payload.get('project_name') or payload.get('name') or '').strip()
+        project_name = fallback_project_name
+        if project_candidate and self._should_use_project_name_candidate(project_candidate, fallback_project_name):
+            project_name = self._humanize_name(project_candidate)
+        repo_candidate = str(payload.get('repo_name') or '').strip()
+        repo_name = self._derive_repo_name(project_name)
+        if repo_candidate and self._should_use_repo_name_candidate(repo_candidate, project_name):
+            repo_name = self._derive_repo_name(repo_candidate)
+        return project_name, repo_name
+
     def _heuristic_fallback(self, prompt_text: str, context: dict | None = None) -> tuple[dict, dict]:
         """Fallback request extraction when Ollama is unavailable."""
         lowered = prompt_text.lower()
@@ -270,6 +491,7 @@ class RequestInterpreter:
             'intent': intent,
             'project_id': matched_project.get('project_id') if matched_project else None,
             'project_name': matched_project.get('name') if matched_project else self._derive_name(prompt_text),
+            'repo_name': None if matched_project else self._derive_repo_name(self._derive_name(prompt_text)),
             'issue_number': issue_number,
             'confidence': 'medium' if matched_project or explicit_new else 'low',
             'reasoning_summary': 'Heuristic routing from chat history and project names.',

ai_software_factory/config.py

@@ -1,12 +1,210 @@
 """Configuration settings for AI Software Factory."""
 
+import json
 import os
 from typing import Optional
 from pathlib import Path
+from urllib.parse import urlparse
 from pydantic import Field
 from pydantic_settings import BaseSettings, SettingsConfigDict
 
 
+def _normalize_service_url(value: str, default_scheme: str = "https") -> str:
+    """Normalize service URLs so host-only values still become valid absolute URLs."""
+    normalized = (value or "").strip().rstrip("/")
+    if not normalized:
+        return ""
+    if "://" not in normalized:
+        normalized = f"{default_scheme}://{normalized}"
+    parsed = urlparse(normalized)
+    if not parsed.scheme or not parsed.netloc:
+        return ""
+    return normalized
+
+
+EDITABLE_LLM_PROMPTS: dict[str, dict[str, str]] = {
+    'LLM_GUARDRAIL_PROMPT': {
+        'label': 'Global Guardrails',
+        'category': 'guardrail',
+        'description': 'Applied to every outbound external LLM call.',
+    },
+    'LLM_REQUEST_INTERPRETER_GUARDRAIL_PROMPT': {
+        'label': 'Request Interpretation Guardrails',
+        'category': 'guardrail',
+        'description': 'Constrains project routing and continuation selection.',
+    },
+    'LLM_CHANGE_SUMMARY_GUARDRAIL_PROMPT': {
+        'label': 'Change Summary Guardrails',
+        'category': 'guardrail',
+        'description': 'Constrains factual delivery summaries.',
+    },
+    'LLM_PROJECT_NAMING_GUARDRAIL_PROMPT': {
+        'label': 'Project Naming Guardrails',
+        'category': 'guardrail',
+        'description': 'Constrains project display names and repo slugs.',
+    },
+    'LLM_PROJECT_NAMING_SYSTEM_PROMPT': {
+        'label': 'Project Naming System Prompt',
+        'category': 'system_prompt',
+        'description': 'Guides the dedicated new-project naming stage.',
+    },
+    'LLM_PROJECT_ID_GUARDRAIL_PROMPT': {
+        'label': 'Project ID Guardrails',
+        'category': 'guardrail',
+        'description': 'Constrains stable project id generation.',
+    },
+    'LLM_PROJECT_ID_SYSTEM_PROMPT': {
+        'label': 'Project ID System Prompt',
+        'category': 'system_prompt',
+        'description': 'Guides the dedicated project id naming stage.',
+    },
+}
+
+EDITABLE_RUNTIME_SETTINGS: dict[str, dict[str, str]] = {
+    'HOME_ASSISTANT_BATTERY_ENTITY_ID': {
+        'label': 'Battery Entity ID',
+        'category': 'home_assistant',
+        'description': 'Home Assistant entity used for battery state-of-charge gating.',
+        'value_type': 'string',
+    },
+    'HOME_ASSISTANT_SURPLUS_ENTITY_ID': {
+        'label': 'Surplus Power Entity ID',
+        'category': 'home_assistant',
+        'description': 'Home Assistant entity used for export or surplus power gating.',
+        'value_type': 'string',
+    },
+    'HOME_ASSISTANT_BATTERY_FULL_THRESHOLD': {
+        'label': 'Battery Full Threshold',
+        'category': 'home_assistant',
+        'description': 'Minimum battery percentage required before queued prompts may run.',
+        'value_type': 'float',
+    },
+    'HOME_ASSISTANT_SURPLUS_THRESHOLD_WATTS': {
+        'label': 'Surplus Threshold Watts',
+        'category': 'home_assistant',
+        'description': 'Minimum surplus/export power required before queued prompts may run.',
+        'value_type': 'float',
+    },
+    'PROMPT_QUEUE_ENABLED': {
+        'label': 'Queue Telegram Prompts',
+        'category': 'prompt_queue',
+        'description': 'When enabled, Telegram prompts are queued and gated instead of processed immediately.',
+        'value_type': 'boolean',
+    },
+    'PROMPT_QUEUE_AUTO_PROCESS': {
+        'label': 'Auto Process Queue',
+        'category': 'prompt_queue',
+        'description': 'Let the background worker drain the queue automatically when the gate is open.',
+        'value_type': 'boolean',
+    },
+    'PROMPT_QUEUE_FORCE_PROCESS': {
+        'label': 'Force Queue Processing',
+        'category': 'prompt_queue',
+        'description': 'Bypass the Home Assistant energy gate for queued prompts.',
+        'value_type': 'boolean',
+    },
+    'PROMPT_QUEUE_POLL_INTERVAL_SECONDS': {
+        'label': 'Queue Poll Interval Seconds',
+        'category': 'prompt_queue',
+        'description': 'Polling interval for the background queue worker.',
+        'value_type': 'integer',
+    },
+    'PROMPT_QUEUE_MAX_BATCH_SIZE': {
+        'label': 'Queue Max Batch Size',
+        'category': 'prompt_queue',
+        'description': 'Maximum number of queued prompts processed in one batch.',
+        'value_type': 'integer',
+    },
+}
+
+
+def _get_persisted_llm_prompt_override(env_key: str) -> str | None:
+    """Load one persisted LLM prompt override from the database when available."""
+    if env_key not in EDITABLE_LLM_PROMPTS:
+        return None
+    try:
+        try:
+            from .database import get_db_sync
+            from .agents.database_manager import DatabaseManager
+        except ImportError:
+            from database import get_db_sync
+            from agents.database_manager import DatabaseManager
+
+        db = get_db_sync()
+        if db is None:
+            return None
+        try:
+            return DatabaseManager(db).get_llm_prompt_override(env_key)
+        finally:
+            db.close()
+    except Exception:
+        return None
+
+
+def _resolve_llm_prompt_value(env_key: str, fallback: str) -> str:
+    """Resolve one editable prompt from DB override first, then environment/defaults."""
+    override = _get_persisted_llm_prompt_override(env_key)
+    if override is not None:
+        return override.strip()
+    return (fallback or '').strip()
+
+
+def _get_persisted_runtime_setting_override(key: str):
+    """Load one persisted runtime-setting override from the database when available."""
+    if key not in EDITABLE_RUNTIME_SETTINGS:
+        return None
+    try:
+        try:
+            from .database import get_db_sync
+            from .agents.database_manager import DatabaseManager
+        except ImportError:
+            from database import get_db_sync
+            from agents.database_manager import DatabaseManager
+
+        db = get_db_sync()
+        if db is None:
+            return None
+        try:
+            return DatabaseManager(db).get_runtime_setting_override(key)
+        finally:
+            db.close()
+    except Exception:
+        return None
+
+
+def _coerce_runtime_setting_value(key: str, value, fallback):
+    """Coerce a persisted runtime setting override into the expected scalar type."""
+    value_type = EDITABLE_RUNTIME_SETTINGS.get(key, {}).get('value_type')
+    if value is None:
+        return fallback
+    if value_type == 'boolean':
+        if isinstance(value, bool):
+            return value
+        normalized = str(value).strip().lower()
+        if normalized in {'1', 'true', 'yes', 'on'}:
+            return True
+        if normalized in {'0', 'false', 'no', 'off'}:
+            return False
+        return bool(fallback)
+    if value_type == 'integer':
+        try:
+            return int(value)
+        except Exception:
+            return int(fallback)
+    if value_type == 'float':
+        try:
+            return float(value)
+        except Exception:
+            return float(fallback)
+    return str(value).strip()
+
+
+def _resolve_runtime_setting_value(key: str, fallback):
+    """Resolve one editable runtime setting from DB override first, then environment/defaults."""
+    override = _get_persisted_runtime_setting_override(key)
+    return _coerce_runtime_setting_value(key, override, fallback)
+
+
 class Settings(BaseSettings):
     """Application settings loaded from environment variables."""
 
@@ -24,6 +222,34 @@ class Settings(BaseSettings):
     # Ollama settings computed from environment
     OLLAMA_URL: str = "http://ollama:11434"
     OLLAMA_MODEL: str = "llama3"
+    LLM_GUARDRAIL_PROMPT: str = (
+        "You are operating inside AI Software Factory. Follow the requested schema exactly, "
+        "treat provided tool outputs as authoritative, and do not invent repositories, issues, pull requests, or delivery facts."
+    )
+    LLM_REQUEST_INTERPRETER_GUARDRAIL_PROMPT: str = (
+        "For routing and request interpretation: never select archived projects, prefer tracked project IDs from tool outputs, and only reference issues that are explicit in the prompt or available tool data."
+    )
+    LLM_CHANGE_SUMMARY_GUARDRAIL_PROMPT: str = (
+        "For summaries: only describe facts present in the provided context and tool outputs. Never claim a repository, commit, or pull request exists unless it is present in the supplied data."
+    )
+    LLM_PROJECT_NAMING_GUARDRAIL_PROMPT: str = (
+        "For project naming: prefer clear, product-like names and repository slugs that match the user's concrete deliverable. Avoid abstract or instructional words such as purpose, project, system, app, tool, platform, solution, new, create, or test unless the request truly centers on that exact noun. Base the name on the actual artifact or workflow being built, and avoid copying sentence fragments from the prompt. Avoid reusing tracked project identities unless the request is clearly asking for an existing project."
+    )
+    LLM_PROJECT_NAMING_SYSTEM_PROMPT: str = (
+        "You name newly requested software projects. Return only JSON with keys project_name, repo_name, and rationale. Project names should be concise human-readable titles based on the real product, artifact, or workflow being created. Repo names should be lowercase kebab-case slugs derived from that title. Never return generic names like purpose, project, system, app, tool, platform, solution, harness, or test by themselves, and never return a repo_name that is a copied sentence fragment from the prompt. Prefer 2 to 4 specific words when possible."
+    )
+    LLM_PROJECT_ID_GUARDRAIL_PROMPT: str = (
+        "For project ids: produce short stable slugs for newly created projects. Avoid collisions with known project ids and keep ids lowercase with hyphens."
+    )
+    LLM_PROJECT_ID_SYSTEM_PROMPT: str = (
+        "You derive stable project ids for new projects. Return only JSON with keys project_id and rationale. project_id must be a short lowercase kebab-case slug without spaces."
+    )
+    LLM_TOOL_ALLOWLIST: str = "gitea_project_catalog,gitea_project_state,gitea_project_issues,gitea_pull_requests"
+    LLM_TOOL_CONTEXT_LIMIT: int = 5
+    LLM_LIVE_TOOL_ALLOWLIST: str = "gitea_lookup_issue,gitea_lookup_pull_request"
+    LLM_LIVE_TOOL_STAGE_ALLOWLIST: str = "request_interpretation,change_summary"
+    LLM_LIVE_TOOL_STAGE_TOOL_MAP: str = ""
+    LLM_MAX_TOOL_CALL_ROUNDS: int = 1
 
     # Gitea settings
     GITEA_URL: str = "https://gitea.yourserver.com"
@@ -47,6 +273,19 @@ class Settings(BaseSettings):
     TELEGRAM_BOT_TOKEN: str = ""
     TELEGRAM_CHAT_ID: str = ""
 
+    # Home Assistant and prompt queue settings
+    HOME_ASSISTANT_URL: str = ""
+    HOME_ASSISTANT_TOKEN: str = ""
+    HOME_ASSISTANT_BATTERY_ENTITY_ID: str = ""
+    HOME_ASSISTANT_SURPLUS_ENTITY_ID: str = ""
+    HOME_ASSISTANT_BATTERY_FULL_THRESHOLD: float = 95.0
+    HOME_ASSISTANT_SURPLUS_THRESHOLD_WATTS: float = 100.0
+    PROMPT_QUEUE_ENABLED: bool = False
+    PROMPT_QUEUE_AUTO_PROCESS: bool = True
+    PROMPT_QUEUE_FORCE_PROCESS: bool = False
+    PROMPT_QUEUE_POLL_INTERVAL_SECONDS: int = 60
+    PROMPT_QUEUE_MAX_BATCH_SIZE: int = 1
+
     # PostgreSQL settings
     POSTGRES_HOST: str = "localhost"
     POSTGRES_PORT: int = 5432
@@ -131,10 +370,138 @@ class Settings(BaseSettings):
         """Get Ollama URL with trimmed whitespace."""
         return self.OLLAMA_URL.strip()
 
+    @property
+    def llm_guardrail_prompt(self) -> str:
+        """Get the global guardrail prompt used for all external LLM calls."""
+        return _resolve_llm_prompt_value('LLM_GUARDRAIL_PROMPT', self.LLM_GUARDRAIL_PROMPT)
+
+    @property
+    def llm_request_interpreter_guardrail_prompt(self) -> str:
+        """Get the request-interpretation specific guardrail prompt."""
+        return _resolve_llm_prompt_value('LLM_REQUEST_INTERPRETER_GUARDRAIL_PROMPT', self.LLM_REQUEST_INTERPRETER_GUARDRAIL_PROMPT)
+
+    @property
+    def llm_change_summary_guardrail_prompt(self) -> str:
+        """Get the change-summary specific guardrail prompt."""
+        return _resolve_llm_prompt_value('LLM_CHANGE_SUMMARY_GUARDRAIL_PROMPT', self.LLM_CHANGE_SUMMARY_GUARDRAIL_PROMPT)
+
+    @property
+    def llm_project_naming_guardrail_prompt(self) -> str:
+        """Get the project-naming specific guardrail prompt."""
+        return _resolve_llm_prompt_value('LLM_PROJECT_NAMING_GUARDRAIL_PROMPT', self.LLM_PROJECT_NAMING_GUARDRAIL_PROMPT)
+
+    @property
+    def llm_project_naming_system_prompt(self) -> str:
+        """Get the project-naming system prompt."""
+        return _resolve_llm_prompt_value('LLM_PROJECT_NAMING_SYSTEM_PROMPT', self.LLM_PROJECT_NAMING_SYSTEM_PROMPT)
+
+    @property
+    def llm_project_id_guardrail_prompt(self) -> str:
+        """Get the project-id naming specific guardrail prompt."""
+        return _resolve_llm_prompt_value('LLM_PROJECT_ID_GUARDRAIL_PROMPT', self.LLM_PROJECT_ID_GUARDRAIL_PROMPT)
+
+    @property
+    def llm_project_id_system_prompt(self) -> str:
+        """Get the project-id naming system prompt."""
+        return _resolve_llm_prompt_value('LLM_PROJECT_ID_SYSTEM_PROMPT', self.LLM_PROJECT_ID_SYSTEM_PROMPT)
+
+    @property
+    def editable_llm_prompts(self) -> list[dict[str, str]]:
+        """Return metadata for all LLM prompts that may be persisted and edited from the UI."""
+        prompts = []
+        for env_key, metadata in EDITABLE_LLM_PROMPTS.items():
+            prompts.append(
+                {
+                    'key': env_key,
+                    'label': metadata['label'],
+                    'category': metadata['category'],
+                    'description': metadata['description'],
+                    'default_value': (getattr(self, env_key, '') or '').strip(),
+                    'value': _resolve_llm_prompt_value(env_key, getattr(self, env_key, '')),
+                }
+            )
+        return prompts
+
+    @property
+    def editable_runtime_settings(self) -> list[dict]:
+        """Return metadata for all DB-editable runtime settings."""
+        items = []
+        for key, metadata in EDITABLE_RUNTIME_SETTINGS.items():
+            default_value = getattr(self, key)
+            value = _resolve_runtime_setting_value(key, default_value)
+            items.append(
+                {
+                    'key': key,
+                    'label': metadata['label'],
+                    'category': metadata['category'],
+                    'description': metadata['description'],
+                    'value_type': metadata['value_type'],
+                    'default_value': default_value,
+                    'value': value,
+                }
+            )
+        return items
+
+    @property
+    def llm_tool_allowlist(self) -> list[str]:
+        """Get the allowed LLM tool names as a normalized list."""
+        return [item.strip() for item in self.LLM_TOOL_ALLOWLIST.split(',') if item.strip()]
+
+    @property
+    def llm_tool_context_limit(self) -> int:
+        """Get the number of items to expose per mediated tool payload."""
+        return max(int(self.LLM_TOOL_CONTEXT_LIMIT), 1)
+
+    @property
+    def llm_live_tool_allowlist(self) -> list[str]:
+        """Get the allowed live tool-call names for model-driven lookup requests."""
+        return [item.strip() for item in self.LLM_LIVE_TOOL_ALLOWLIST.split(',') if item.strip()]
+
+    @property
+    def llm_live_tool_stage_allowlist(self) -> list[str]:
+        """Get the LLM stages where live tool requests are enabled."""
+        return [item.strip() for item in self.LLM_LIVE_TOOL_STAGE_ALLOWLIST.split(',') if item.strip()]
+
+    @property
+    def llm_live_tool_stage_tool_map(self) -> dict[str, list[str]]:
+        """Get an optional per-stage live tool map that overrides the simple stage allowlist."""
+        raw = (self.LLM_LIVE_TOOL_STAGE_TOOL_MAP or '').strip()
+        if not raw:
+            return {}
+        try:
+            parsed = json.loads(raw)
+        except Exception:
+            return {}
+        if not isinstance(parsed, dict):
+            return {}
+        allowed_tools = set(self.llm_live_tool_allowlist)
+        normalized: dict[str, list[str]] = {}
+        for stage, tools in parsed.items():
+            if not isinstance(stage, str):
+                continue
+            if not isinstance(tools, list):
+                continue
+            normalized[stage.strip()] = [str(tool).strip() for tool in tools if str(tool).strip() in allowed_tools]
+        return normalized
+
+    def llm_live_tools_for_stage(self, stage: str) -> list[str]:
+        """Return live tools enabled for a specific LLM stage."""
+        stage_map = self.llm_live_tool_stage_tool_map
+        if stage_map:
+            return stage_map.get(stage, [])
+        if stage not in set(self.llm_live_tool_stage_allowlist):
+            return []
+        return self.llm_live_tool_allowlist
+
+    @property
+    def llm_max_tool_call_rounds(self) -> int:
+        """Get the maximum number of model-driven live tool-call rounds per LLM request."""
+        return max(int(self.LLM_MAX_TOOL_CALL_ROUNDS), 0)
+
     @property
     def gitea_url(self) -> str:
         """Get Gitea URL with trimmed whitespace."""
-        return self.GITEA_URL.strip()
+        return _normalize_service_url(self.GITEA_URL)
 
     @property
     def gitea_token(self) -> str:
@@ -159,12 +526,12 @@ class Settings(BaseSettings):
     @property
     def n8n_webhook_url(self) -> str:
         """Get n8n webhook URL with trimmed whitespace."""
-        return self.N8N_WEBHOOK_URL.strip()
+        return _normalize_service_url(self.N8N_WEBHOOK_URL, default_scheme="http")
 
     @property
     def n8n_api_url(self) -> str:
         """Get n8n API URL with trimmed whitespace."""
-        return self.N8N_API_URL.strip()
+        return _normalize_service_url(self.N8N_API_URL, default_scheme="http")
 
     @property
     def n8n_api_key(self) -> str:
@@ -189,7 +556,62 @@ class Settings(BaseSettings):
     @property
     def backend_public_url(self) -> str:
         """Get backend public URL with trimmed whitespace."""
-        return self.BACKEND_PUBLIC_URL.strip().rstrip("/")
+        return _normalize_service_url(self.BACKEND_PUBLIC_URL, default_scheme="http")
+
+    @property
+    def home_assistant_url(self) -> str:
+        """Get Home Assistant URL with trimmed whitespace."""
+        return _normalize_service_url(self.HOME_ASSISTANT_URL, default_scheme="http")
+
+    @property
+    def home_assistant_token(self) -> str:
+        """Get Home Assistant token with trimmed whitespace."""
+        return self.HOME_ASSISTANT_TOKEN.strip()
+
+    @property
+    def home_assistant_battery_entity_id(self) -> str:
+        """Get the Home Assistant battery state entity id."""
+        return str(_resolve_runtime_setting_value('HOME_ASSISTANT_BATTERY_ENTITY_ID', self.HOME_ASSISTANT_BATTERY_ENTITY_ID)).strip()
+
+    @property
+    def home_assistant_surplus_entity_id(self) -> str:
+        """Get the Home Assistant surplus power entity id."""
+        return str(_resolve_runtime_setting_value('HOME_ASSISTANT_SURPLUS_ENTITY_ID', self.HOME_ASSISTANT_SURPLUS_ENTITY_ID)).strip()
+
+    @property
+    def home_assistant_battery_full_threshold(self) -> float:
+        """Get the minimum battery SoC percentage for queue processing."""
+        return float(_resolve_runtime_setting_value('HOME_ASSISTANT_BATTERY_FULL_THRESHOLD', self.HOME_ASSISTANT_BATTERY_FULL_THRESHOLD))
+
+    @property
+    def home_assistant_surplus_threshold_watts(self) -> float:
+        """Get the minimum export/surplus power threshold for queue processing."""
+        return float(_resolve_runtime_setting_value('HOME_ASSISTANT_SURPLUS_THRESHOLD_WATTS', self.HOME_ASSISTANT_SURPLUS_THRESHOLD_WATTS))
+
+    @property
+    def prompt_queue_enabled(self) -> bool:
+        """Whether Telegram prompts should be queued instead of processed immediately."""
+        return bool(_resolve_runtime_setting_value('PROMPT_QUEUE_ENABLED', self.PROMPT_QUEUE_ENABLED))
+
+    @property
+    def prompt_queue_auto_process(self) -> bool:
+        """Whether the background worker should automatically process queued prompts."""
+        return bool(_resolve_runtime_setting_value('PROMPT_QUEUE_AUTO_PROCESS', self.PROMPT_QUEUE_AUTO_PROCESS))
+
+    @property
+    def prompt_queue_force_process(self) -> bool:
+        """Whether queued prompts should bypass the Home Assistant energy gate."""
+        return bool(_resolve_runtime_setting_value('PROMPT_QUEUE_FORCE_PROCESS', self.PROMPT_QUEUE_FORCE_PROCESS))
+
+    @property
+    def prompt_queue_poll_interval_seconds(self) -> int:
+        """Get the queue polling interval for background processing."""
+        return max(int(_resolve_runtime_setting_value('PROMPT_QUEUE_POLL_INTERVAL_SECONDS', self.PROMPT_QUEUE_POLL_INTERVAL_SECONDS)), 5)
+
+    @property
+    def prompt_queue_max_batch_size(self) -> int:
+        """Get the maximum number of queued prompts to process in one batch."""
+        return max(int(_resolve_runtime_setting_value('PROMPT_QUEUE_MAX_BATCH_SIZE', self.PROMPT_QUEUE_MAX_BATCH_SIZE)), 1)
 
     @property
     def projects_root(self) -> Path:

ai_software_factory/database.py

@@ -6,7 +6,7 @@ from urllib.parse import urlparse
 
 from alembic import command
 from alembic.config import Config
-from sqlalchemy import create_engine, event, text
+from sqlalchemy import create_engine, text
 from sqlalchemy.engine import Engine
 from sqlalchemy.orm import Session, sessionmaker
 
@@ -64,20 +64,6 @@ def get_engine() -> Engine:
             pool_timeout=settings.DB_POOL_TIMEOUT or 30
         )
     
-    # Event listener for connection checkout (PostgreSQL only)
-    if not settings.use_sqlite:
-        @event.listens_for(engine, "checkout")
-        def receive_checkout(dbapi_connection, connection_record, connection_proxy):
-            """Log connection checkout for audit purposes."""
-            if settings.LOG_LEVEL in ("DEBUG", "INFO"):
-                print(f"DB Connection checked out from pool")
-        
-        @event.listens_for(engine, "checkin")
-        def receive_checkin(dbapi_connection, connection_record):
-            """Log connection checkin for audit purposes."""
-            if settings.LOG_LEVEL == "DEBUG":
-                print(f"DB Connection returned to pool")
-    
     return engine
 
 

ai_software_factory/main.py

@@ -13,6 +13,7 @@ The NiceGUI frontend provides:
 
 from __future__ import annotations
 
+import asyncio
 from contextlib import asynccontextmanager
 import json
 import re
@@ -29,7 +30,9 @@ try:
     from . import database as database_module
     from .agents.change_summary import ChangeSummaryGenerator
     from .agents.database_manager import DatabaseManager
+    from .agents.home_assistant import HomeAssistantAgent
     from .agents.request_interpreter import RequestInterpreter
+    from .agents.llm_service import LLMServiceClient
     from .agents.orchestrator import AgentOrchestrator
     from .agents.n8n_setup import N8NSetupAgent
     from .agents.prompt_workflow import PromptWorkflowManager
@@ -40,7 +43,9 @@ except ImportError:
     import database as database_module
     from agents.change_summary import ChangeSummaryGenerator
     from agents.database_manager import DatabaseManager
+    from agents.home_assistant import HomeAssistantAgent
     from agents.request_interpreter import RequestInterpreter
+    from agents.llm_service import LLMServiceClient
     from agents.orchestrator import AgentOrchestrator
     from agents.n8n_setup import N8NSetupAgent
     from agents.prompt_workflow import PromptWorkflowManager
@@ -57,7 +62,16 @@ async def lifespan(_app: FastAPI):
     print(
         f"Runtime configuration: database_backend={runtime['backend']} target={runtime['target']}"
     )
+    queue_worker = asyncio.create_task(_prompt_queue_worker())
+    try:
         yield
+    finally:
+        if queue_worker is not None:
+            queue_worker.cancel()
+            try:
+                await queue_worker
+            except asyncio.CancelledError:
+                pass
 
 
 app = FastAPI(lifespan=lifespan)
@@ -92,6 +106,26 @@ class FreeformSoftwareRequest(BaseModel):
     source: str = 'telegram'
     chat_id: str | None = None
     chat_type: str | None = None
+    process_now: bool = False
+
+
+class PromptQueueProcessRequest(BaseModel):
+    """Request body for manual queue processing."""
+
+    force: bool = False
+    limit: int = Field(default=1, ge=1, le=25)
+
+
+class LLMPromptSettingUpdateRequest(BaseModel):
+    """Request body for persisting one editable LLM prompt override."""
+
+    value: str = Field(default='')
+
+
+class RuntimeSettingUpdateRequest(BaseModel):
+    """Request body for persisting one editable runtime setting override."""
+
+    value: str | bool | int | float | None = None
 
 
 class GiteaRepositoryOnboardRequest(BaseModel):
@@ -109,6 +143,75 @@ def _build_project_id(name: str) -> str:
     return f"{slug}-{uuid4().hex[:8]}"
 
 
+def _build_project_slug(name: str) -> str:
+    """Normalize a project name into a kebab-case identifier slug."""
+    return PROJECT_ID_PATTERN.sub("-", name.strip().lower()).strip("-") or "project"
+
+
+def _ensure_unique_identifier(base_slug: str, reserved_ids: set[str]) -> str:
+    """Return a unique identifier using deterministic numeric suffixes when needed."""
+    normalized = _build_project_slug(base_slug)
+    if normalized not in reserved_ids:
+        return normalized
+    suffix = 2
+    while f"{normalized}-{suffix}" in reserved_ids:
+        suffix += 1
+    return f"{normalized}-{suffix}"
+
+
+def _build_project_identity_context(manager: DatabaseManager) -> list[dict]:
+    """Build a compact project catalog for naming stages."""
+    projects = []
+    for history in manager.get_all_projects(include_archived=True):
+        repository = manager._get_project_repository(history) or {}
+        projects.append(
+            {
+                'project_id': history.project_id,
+                'name': history.project_name,
+                'description': history.description,
+                'repository': {
+                    'owner': repository.get('owner'),
+                    'name': repository.get('name'),
+                },
+            }
+        )
+    return projects
+
+
+async def _derive_project_id_for_request(
+    request: SoftwareRequest,
+    *,
+    prompt_text: str,
+    prompt_routing: dict | None,
+    existing_projects: list[dict],
+) -> tuple[str, dict | None]:
+    """Derive a stable project id for a newly created project."""
+    reserved_ids = {str(project.get('project_id')).strip() for project in existing_projects if project.get('project_id')}
+    fallback_id = _ensure_unique_identifier((prompt_routing or {}).get('project_name') or request.name, reserved_ids)
+    user_prompt = (
+        f"Original user prompt:\n{prompt_text}\n\n"
+        f"Structured request:\n{json.dumps({'name': request.name, 'description': request.description, 'features': request.features, 'tech_stack': request.tech_stack}, indent=2)}\n\n"
+        f"Naming context:\n{json.dumps(prompt_routing or {}, indent=2)}\n\n"
+        f"Reserved project ids:\n{json.dumps(sorted(reserved_ids))}\n\n"
+        "Suggest the best stable project id for this new project."
+    )
+    content, trace = await LLMServiceClient().chat_with_trace(
+        stage='project_id_naming',
+        system_prompt=database_module.settings.llm_project_id_system_prompt,
+        user_prompt=user_prompt,
+        tool_context_input={'projects': existing_projects},
+        expect_json=True,
+    )
+    if content:
+        try:
+            parsed = json.loads(content)
+            candidate = parsed.get('project_id') or parsed.get('slug') or request.name
+            return _ensure_unique_identifier(str(candidate), reserved_ids), trace
+        except Exception:
+            pass
+    return fallback_id, trace
+
+
 def _serialize_project(history: ProjectHistory) -> dict:
     """Serialize a project history row for API responses."""
     return {
@@ -176,13 +279,15 @@ async def _run_generation(
     prompt_source_context: dict | None = None,
     prompt_routing: dict | None = None,
     preferred_project_id: str | None = None,
+    repo_name_override: str | None = None,
     related_issue: dict | None = None,
 ) -> dict:
     """Run the shared generation pipeline for a structured request."""
     database_module.init_db()
 
     manager = DatabaseManager(db)
-    reusable_history = manager.get_project_by_id(preferred_project_id, include_archived=False) if preferred_project_id else manager.get_latest_project_by_name(request.name)
+    is_explicit_new_project = (prompt_routing or {}).get('intent') == 'new_project'
+    reusable_history = manager.get_project_by_id(preferred_project_id, include_archived=False) if preferred_project_id else (None if is_explicit_new_project else manager.get_latest_project_by_name(request.name))
     if reusable_history and database_module.settings.gitea_url and database_module.settings.gitea_token:
         try:
             from .agents.gitea import GiteaAPI
@@ -197,14 +302,23 @@ async def _run_generation(
             ),
             project_id=reusable_history.project_id,
         )
+    project_id_trace = None
+    resolved_prompt_text = prompt_text or _compose_prompt_text(request)
     if preferred_project_id and reusable_history is not None:
         project_id = reusable_history.project_id
-    elif reusable_history and manager.get_open_pull_request(project_id=reusable_history.project_id):
+    elif reusable_history and not is_explicit_new_project:
         project_id = reusable_history.project_id
+    else:
+        if is_explicit_new_project or prompt_text:
+            project_id, project_id_trace = await _derive_project_id_for_request(
+                request,
+                prompt_text=resolved_prompt_text,
+                prompt_routing=prompt_routing,
+                existing_projects=_build_project_identity_context(manager),
+            )
         else:
             project_id = _build_project_id(request.name)
         reusable_history = None
-    resolved_prompt_text = prompt_text or _compose_prompt_text(request)
     orchestrator = AgentOrchestrator(
         project_id=project_id,
         project_name=request.name,
@@ -217,6 +331,7 @@ async def _run_generation(
         existing_history=reusable_history,
         prompt_source_context=prompt_source_context,
         prompt_routing=prompt_routing,
+        repo_name_override=repo_name_override,
         related_issue_hint=related_issue,
     )
     result = await orchestrator.run()
@@ -233,6 +348,8 @@ async def _run_generation(
     response_data = _serialize_project(history)
     response_data['logs'] = [_serialize_project_log(log) for log in project_logs]
     response_data['ui_data'] = result.get('ui_data')
+    response_data['generation_debug'] = ((result.get('ui_data') or {}).get('generation_debug'))
+    response_data['git_debug'] = ((result.get('ui_data') or {}).get('git'))
     response_data['features'] = request.features
     response_data['tech_stack'] = request.tech_stack
     response_data['project_root'] = result.get('project_root', str(_project_root(project_id)))
@@ -240,6 +357,20 @@ async def _run_generation(
     response_data['repository'] = result.get('repository')
     response_data['related_issue'] = result.get('related_issue') or (result.get('ui_data') or {}).get('related_issue')
     response_data['pull_request'] = result.get('pull_request') or manager.get_open_pull_request(project_id=project_id)
+    if project_id_trace:
+        manager.log_llm_trace(
+            project_id=project_id,
+            history_id=history.id if history else None,
+            prompt_id=orchestrator.prompt_audit.id if orchestrator.prompt_audit else None,
+            stage=project_id_trace['stage'],
+            provider=project_id_trace['provider'],
+            model=project_id_trace['model'],
+            system_prompt=project_id_trace['system_prompt'],
+            user_prompt=project_id_trace['user_prompt'],
+            assistant_response=project_id_trace['assistant_response'],
+            raw_response=project_id_trace.get('raw_response'),
+            fallback_used=project_id_trace.get('fallback_used', False),
+        )
     summary_context = {
         'name': response_data['name'],
         'description': response_data['description'],
@@ -275,8 +406,18 @@ async def _run_generation(
             fallback_used=summary_trace.get('fallback_used', False),
         )
     response_data['summary_message'] = summary_message
+    response_data['summary_metadata'] = {
+        'provider': summary_trace.get('provider'),
+        'model': summary_trace.get('model'),
+        'fallback_used': bool(summary_trace.get('fallback_used')),
+    }
     response_data['pull_request'] = result.get('pull_request') or manager.get_open_pull_request(project_id=project_id)
-    return {'status': result['status'], 'data': response_data, 'summary_message': summary_message}
+    return {
+        'status': result['status'],
+        'data': response_data,
+        'summary_message': summary_message,
+        'summary_metadata': response_data['summary_metadata'],
+    }
 
 
 def _project_root(project_id: str) -> Path:
@@ -300,6 +441,276 @@ def _create_gitea_api():
     )
 
 
+def _create_home_assistant_agent() -> HomeAssistantAgent:
+    """Create a configured Home Assistant client."""
+    return HomeAssistantAgent(
+        base_url=database_module.settings.home_assistant_url,
+        token=database_module.settings.home_assistant_token,
+    )
+
+
+def _get_gitea_health() -> dict:
+    """Return current Gitea connectivity diagnostics."""
+    if not database_module.settings.gitea_url:
+        return {
+            'status': 'error',
+            'message': 'Gitea URL is not configured.',
+            'base_url': '',
+            'configured': False,
+            'checks': [],
+        }
+    if not database_module.settings.gitea_token:
+        return {
+            'status': 'error',
+            'message': 'Gitea token is not configured.',
+            'base_url': database_module.settings.gitea_url,
+            'configured': False,
+            'checks': [],
+        }
+    response = _create_gitea_api().get_current_user_sync()
+    if response.get('error'):
+        return {
+            'status': 'error',
+            'message': response.get('error'),
+            'base_url': database_module.settings.gitea_url,
+            'configured': True,
+            'checks': [
+                {
+                    'name': 'token_auth',
+                    'ok': False,
+                    'message': response.get('error'),
+                    'url': f"{database_module.settings.gitea_url}/api/v1/user",
+                    'status_code': response.get('status_code'),
+                }
+            ],
+        }
+    username = response.get('login') or response.get('username') or response.get('full_name') or 'unknown'
+    return {
+        'status': 'success',
+        'message': f'Authenticated as {username}.',
+        'base_url': database_module.settings.gitea_url,
+        'configured': True,
+        'checks': [
+            {
+                'name': 'token_auth',
+                'ok': True,
+                'message': f'Authenticated as {username}',
+                'url': f"{database_module.settings.gitea_url}/api/v1/user",
+            }
+        ],
+        'user': username,
+    }
+
+
+def _get_home_assistant_health() -> dict:
+    """Return current Home Assistant connectivity diagnostics."""
+    return _create_home_assistant_agent().health_check_sync()
+
+
+async def _get_queue_gate_status(force: bool = False) -> dict:
+    """Return whether queued prompts may be processed now."""
+    if not database_module.settings.prompt_queue_enabled:
+        return {
+            'status': 'disabled',
+            'allowed': True,
+            'forced': False,
+            'reason': 'Prompt queue is disabled',
+        }
+    if not database_module.settings.home_assistant_url:
+        if force or database_module.settings.prompt_queue_force_process:
+            return {
+                'status': 'success',
+                'allowed': True,
+                'forced': True,
+                'reason': 'Queue override is enabled',
+            }
+        return {
+            'status': 'blocked',
+            'allowed': False,
+            'forced': False,
+            'reason': 'Home Assistant URL is not configured',
+        }
+    return await _create_home_assistant_agent().queue_gate_status(force=force)
+
+
+async def _interpret_freeform_request(request: FreeformSoftwareRequest, manager: DatabaseManager) -> tuple[SoftwareRequest, dict, dict]:
+    """Interpret a free-form request and return the structured request plus routing trace."""
+    interpreter_context = manager.get_interpreter_context(chat_id=request.chat_id, source=request.source)
+    interpreted, interpretation_trace = await RequestInterpreter().interpret_with_trace(
+        request.prompt_text,
+        context=interpreter_context,
+    )
+    routing = interpretation_trace.get('routing') or {}
+    selected_history = manager.get_project_by_id(routing.get('project_id'), include_archived=False) if routing.get('project_id') else None
+    if selected_history is not None and routing.get('intent') != 'new_project':
+        interpreted['name'] = selected_history.project_name
+        interpreted['description'] = selected_history.description or interpreted['description']
+    return SoftwareRequest(**interpreted), routing, interpretation_trace
+
+
+async def _run_freeform_generation(
+    request: FreeformSoftwareRequest,
+    db: Session,
+    *,
+    queue_item_id: int | None = None,
+) -> dict:
+    """Shared free-form request flow used by direct calls and queued processing."""
+    manager = DatabaseManager(db)
+    try:
+        structured_request, routing, interpretation_trace = await _interpret_freeform_request(request, manager)
+        response = await _run_generation(
+            structured_request,
+            db,
+            prompt_text=request.prompt_text,
+            prompt_actor=request.source,
+            prompt_source_context={
+                'chat_id': request.chat_id,
+                'chat_type': request.chat_type,
+                'queue_item_id': queue_item_id,
+            },
+            prompt_routing=routing,
+            preferred_project_id=routing.get('project_id') if routing.get('intent') != 'new_project' else None,
+            repo_name_override=routing.get('repo_name') if routing.get('intent') == 'new_project' else None,
+            related_issue={'number': routing.get('issue_number')} if routing.get('issue_number') is not None else None,
+        )
+        project_data = response.get('data', {})
+        if project_data.get('history_id') is not None:
+            manager = DatabaseManager(db)
+            prompts = manager.get_prompt_events(project_id=project_data.get('project_id'))
+            prompt_id = prompts[0]['id'] if prompts else None
+            manager.log_llm_trace(
+                project_id=project_data.get('project_id'),
+                history_id=project_data.get('history_id'),
+                prompt_id=prompt_id,
+                stage=interpretation_trace['stage'],
+                provider=interpretation_trace['provider'],
+                model=interpretation_trace['model'],
+                system_prompt=interpretation_trace['system_prompt'],
+                user_prompt=interpretation_trace['user_prompt'],
+                assistant_response=interpretation_trace['assistant_response'],
+                raw_response=interpretation_trace.get('raw_response'),
+                fallback_used=interpretation_trace.get('fallback_used', False),
+            )
+            naming_trace = interpretation_trace.get('project_naming')
+            if naming_trace:
+                manager.log_llm_trace(
+                    project_id=project_data.get('project_id'),
+                    history_id=project_data.get('history_id'),
+                    prompt_id=prompt_id,
+                    stage=naming_trace['stage'],
+                    provider=naming_trace['provider'],
+                    model=naming_trace['model'],
+                    system_prompt=naming_trace['system_prompt'],
+                    user_prompt=naming_trace['user_prompt'],
+                    assistant_response=naming_trace['assistant_response'],
+                    raw_response=naming_trace.get('raw_response'),
+                    fallback_used=naming_trace.get('fallback_used', False),
+                )
+        response['interpreted_request'] = structured_request.model_dump()
+        response['routing'] = routing
+        response['llm_trace'] = interpretation_trace
+        response['source'] = {
+            'type': request.source,
+            'chat_id': request.chat_id,
+            'chat_type': request.chat_type,
+        }
+        if queue_item_id is not None:
+            DatabaseManager(db).complete_queued_prompt(
+                queue_item_id,
+                {
+                    'project_id': project_data.get('project_id'),
+                    'history_id': project_data.get('history_id'),
+                    'status': response.get('status'),
+                },
+            )
+        return response
+    except Exception as exc:
+        if queue_item_id is not None:
+            DatabaseManager(db).fail_queued_prompt(queue_item_id, str(exc))
+        raise
+
+
+async def _process_prompt_queue_batch(limit: int = 1, force: bool = False) -> dict:
+    """Process up to `limit` queued prompts if the energy gate allows it."""
+    queue_gate = await _get_queue_gate_status(force=force)
+    if not queue_gate.get('allowed'):
+        db = database_module.get_db_sync()
+        try:
+            summary = DatabaseManager(db).get_prompt_queue_summary()
+        finally:
+            db.close()
+        return {
+            'status': queue_gate.get('status', 'blocked'),
+            'processed_count': 0,
+            'queue_gate': queue_gate,
+            'queue_summary': summary,
+            'processed': [],
+        }
+
+    processed = []
+    for _ in range(max(limit, 1)):
+        claim_db = database_module.get_db_sync()
+        try:
+            claimed = DatabaseManager(claim_db).claim_next_queued_prompt()
+        finally:
+            claim_db.close()
+        if claimed is None:
+            break
+        work_db = database_module.get_db_sync()
+        try:
+            request = FreeformSoftwareRequest(
+                prompt_text=claimed['prompt_text'],
+                source=claimed['source'] or 'telegram',
+                chat_id=claimed.get('chat_id'),
+                chat_type=claimed.get('chat_type'),
+                process_now=True,
+            )
+            response = await _run_freeform_generation(request, work_db, queue_item_id=claimed['id'])
+            processed.append(
+                {
+                    'queue_item_id': claimed['id'],
+                    'project_id': (response.get('data') or {}).get('project_id'),
+                    'status': response.get('status'),
+                }
+            )
+        except Exception as exc:
+            DatabaseManager(work_db).fail_queued_prompt(claimed['id'], str(exc))
+            processed.append({'queue_item_id': claimed['id'], 'status': 'failed', 'error': str(exc)})
+        finally:
+            work_db.close()
+
+    summary_db = database_module.get_db_sync()
+    try:
+        summary = DatabaseManager(summary_db).get_prompt_queue_summary()
+    finally:
+        summary_db.close()
+    return {
+        'status': 'success',
+        'processed_count': len(processed),
+        'processed': processed,
+        'queue_gate': queue_gate,
+        'queue_summary': summary,
+    }
+
+
+async def _prompt_queue_worker() -> None:
+    """Background worker that drains the prompt queue when the energy gate opens."""
+    while True:
+        try:
+            if database_module.settings.prompt_queue_enabled and database_module.settings.prompt_queue_auto_process:
+                await _process_prompt_queue_batch(
+                    limit=database_module.settings.prompt_queue_max_batch_size,
+                    force=database_module.settings.prompt_queue_force_process,
+                )
+        except Exception as exc:
+            db = database_module.get_db_sync()
+            try:
+                DatabaseManager(db).log_system_event('prompt-queue', 'ERROR', f'Queue worker error: {exc}')
+            finally:
+                db.close()
+        await asyncio.sleep(database_module.settings.prompt_queue_poll_interval_seconds)
+
+
 def _resolve_n8n_api_url(explicit_url: str | None = None) -> str:
     """Resolve the effective n8n API URL from explicit input or settings."""
     if explicit_url and explicit_url.strip():
@@ -322,8 +733,15 @@ def read_api_info():
             '/',
             '/api',
             '/health',
+            '/llm/runtime',
+            '/llm/prompts',
+            '/llm/prompts/{prompt_key}',
+            '/settings/runtime',
+            '/settings/runtime/{setting_key}',
             '/generate',
             '/generate/text',
+            '/queue',
+            '/queue/process',
             '/projects',
             '/status/{project_id}',
             '/audit/projects',
@@ -344,7 +762,9 @@ def read_api_info():
             '/projects/{project_id}/prompts/{prompt_id}/undo',
             '/projects/{project_id}/sync-repository',
             '/gitea/repos',
+            '/gitea/health',
             '/gitea/repos/onboard',
+            '/home-assistant/health',
             '/n8n/health',
             '/n8n/setup',
         ],
@@ -355,14 +775,91 @@ def read_api_info():
 def health_check():
     """Health check endpoint."""
     runtime = database_module.get_database_runtime_summary()
+    queue_summary = {'queued': 0, 'processing': 0, 'completed': 0, 'failed': 0, 'total': 0, 'next_item': None}
+    db = database_module.get_db_sync()
+    try:
+        try:
+            queue_summary = DatabaseManager(db).get_prompt_queue_summary()
+        except Exception:
+            pass
+    finally:
+        db.close()
     return {
         'status': 'healthy',
         'database': runtime['backend'],
         'database_target': runtime['target'],
         'database_name': runtime['database'],
+        'integrations': {
+            'gitea': _get_gitea_health(),
+            'home_assistant': _get_home_assistant_health(),
+        },
+        'prompt_queue': {
+            'enabled': database_module.settings.prompt_queue_enabled,
+            'auto_process': database_module.settings.prompt_queue_auto_process,
+            'force_process': database_module.settings.prompt_queue_force_process,
+            'summary': queue_summary,
+        },
     }
 
 
+@app.get('/llm/runtime')
+def get_llm_runtime():
+    """Return the active external LLM runtime, guardrail, and tool configuration."""
+    return LLMServiceClient().get_runtime_configuration()
+
+
+@app.get('/llm/prompts')
+def get_llm_prompt_settings(db: DbSession):
+    """Return editable LLM prompt settings with DB overrides merged over environment defaults."""
+    return {'prompts': DatabaseManager(db).get_llm_prompt_settings()}
+
+
+@app.put('/llm/prompts/{prompt_key}')
+def update_llm_prompt_setting(prompt_key: str, request: LLMPromptSettingUpdateRequest, db: DbSession):
+    """Persist one editable LLM prompt override into the database."""
+    database_module.init_db()
+    result = DatabaseManager(db).save_llm_prompt_setting(prompt_key, request.value, actor='api')
+    if result.get('status') == 'error':
+        raise HTTPException(status_code=400, detail=result.get('message', 'Prompt save failed'))
+    return result
+
+
+@app.delete('/llm/prompts/{prompt_key}')
+def reset_llm_prompt_setting(prompt_key: str, db: DbSession):
+    """Reset one editable LLM prompt override back to the environment/default value."""
+    database_module.init_db()
+    result = DatabaseManager(db).reset_llm_prompt_setting(prompt_key, actor='api')
+    if result.get('status') == 'error':
+        raise HTTPException(status_code=400, detail=result.get('message', 'Prompt reset failed'))
+    return result
+
+
+@app.get('/settings/runtime')
+def get_runtime_settings(db: DbSession):
+    """Return editable runtime settings with DB overrides merged over environment defaults."""
+    return {'settings': DatabaseManager(db).get_runtime_settings()}
+
+
+@app.put('/settings/runtime/{setting_key}')
+def update_runtime_setting(setting_key: str, request: RuntimeSettingUpdateRequest, db: DbSession):
+    """Persist one editable runtime setting override into the database."""
+    database_module.init_db()
+    result = DatabaseManager(db).save_runtime_setting(setting_key, request.value, actor='api')
+    if result.get('status') == 'error':
+        raise HTTPException(status_code=400, detail=result.get('message', 'Runtime setting save failed'))
+    return result
+
+
+@app.delete('/settings/runtime/{setting_key}')
+def reset_runtime_setting(setting_key: str, db: DbSession):
+    """Reset one editable runtime setting override back to the environment/default value."""
+    database_module.init_db()
+    result = DatabaseManager(db).reset_runtime_setting(setting_key, actor='api')
+    if result.get('status') == 'error':
+        raise HTTPException(status_code=400, detail=result.get('message', 'Runtime setting reset failed'))
+    return result
+
+
 @app.post('/generate')
 async def generate_software(request: SoftwareRequest, db: DbSession):
     """Create and record a software-generation request."""
@@ -388,58 +885,64 @@ async def generate_software_from_text(request: FreeformSoftwareRequest, db: DbSe
             },
         }
 
+    if request.source == 'telegram' and database_module.settings.prompt_queue_enabled and not request.process_now:
         manager = DatabaseManager(db)
-    interpreter_context = manager.get_interpreter_context(chat_id=request.chat_id, source=request.source)
-    interpreted, interpretation_trace = await RequestInterpreter().interpret_with_trace(
-        request.prompt_text,
-        context=interpreter_context,
-    )
-    routing = interpretation_trace.get('routing') or {}
-    selected_history = manager.get_project_by_id(routing.get('project_id'), include_archived=False) if routing.get('project_id') else None
-    if selected_history is not None and routing.get('intent') != 'new_project':
-        interpreted['name'] = selected_history.project_name
-        interpreted['description'] = selected_history.description or interpreted['description']
-    structured_request = SoftwareRequest(**interpreted)
-    response = await _run_generation(
-        structured_request,
-        db,
+        queue_item = manager.enqueue_prompt(
             prompt_text=request.prompt_text,
-        prompt_actor=request.source,
-        prompt_source_context={
-            'chat_id': request.chat_id,
-            'chat_type': request.chat_type,
-        },
-        prompt_routing=routing,
-        preferred_project_id=routing.get('project_id') if routing.get('intent') != 'new_project' else None,
-        related_issue={'number': routing.get('issue_number')} if routing.get('issue_number') is not None else None,
+            source=request.source,
+            chat_id=request.chat_id,
+            chat_type=request.chat_type,
+            source_context={'chat_id': request.chat_id, 'chat_type': request.chat_type},
         )
-    project_data = response.get('data', {})
-    if project_data.get('history_id') is not None:
-        manager = DatabaseManager(db)
-        prompts = manager.get_prompt_events(project_id=project_data.get('project_id'))
-        prompt_id = prompts[0]['id'] if prompts else None
-        manager.log_llm_trace(
-            project_id=project_data.get('project_id'),
-            history_id=project_data.get('history_id'),
-            prompt_id=prompt_id,
-            stage=interpretation_trace['stage'],
-            provider=interpretation_trace['provider'],
-            model=interpretation_trace['model'],
-            system_prompt=interpretation_trace['system_prompt'],
-            user_prompt=interpretation_trace['user_prompt'],
-            assistant_response=interpretation_trace['assistant_response'],
-            raw_response=interpretation_trace.get('raw_response'),
-            fallback_used=interpretation_trace.get('fallback_used', False),
-        )
-    response['interpreted_request'] = interpreted
-    response['routing'] = routing
-    response['llm_trace'] = interpretation_trace
-    response['source'] = {
+        return {
+            'status': 'queued',
+            'message': 'Prompt queued for energy-aware processing.',
+            'queue_item': queue_item,
+            'queue_summary': manager.get_prompt_queue_summary(),
+            'queue_gate': await _get_queue_gate_status(force=False),
+            'source': {
                 'type': request.source,
                 'chat_id': request.chat_id,
                 'chat_type': request.chat_type,
+            },
         }
-    return response
+
+    return await _run_freeform_generation(request, db)
+
+
+@app.get('/queue')
+def get_prompt_queue(db: DbSession):
+    """Return queued prompt items and prompt queue configuration."""
+    manager = DatabaseManager(db)
+    return {
+        'queue': manager.get_prompt_queue(),
+        'summary': manager.get_prompt_queue_summary(),
+        'config': {
+            'enabled': database_module.settings.prompt_queue_enabled,
+            'auto_process': database_module.settings.prompt_queue_auto_process,
+            'force_process': database_module.settings.prompt_queue_force_process,
+            'poll_interval_seconds': database_module.settings.prompt_queue_poll_interval_seconds,
+            'max_batch_size': database_module.settings.prompt_queue_max_batch_size,
+        },
+    }
+
+
+@app.post('/queue/process')
+async def process_prompt_queue(request: PromptQueueProcessRequest):
+    """Manually process queued prompts, optionally bypassing the HA gate."""
+    return await _process_prompt_queue_batch(limit=request.limit, force=request.force)
+
+
+@app.get('/gitea/health')
+def get_gitea_health():
+    """Return Gitea integration connectivity diagnostics."""
+    return _get_gitea_health()
+
+
+@app.get('/home-assistant/health')
+def get_home_assistant_health():
+    """Return Home Assistant integration connectivity diagnostics."""
+    return _get_home_assistant_health()
 
 
 @app.get('/projects')
@@ -623,13 +1126,18 @@ def delete_project(project_id: str, db: DbSession):
     remote_delete = None
     if repository and repository.get('mode') != 'shared' and repository.get('owner') and repository.get('name') and database_module.settings.gitea_url and database_module.settings.gitea_token:
         remote_delete = _create_gitea_api().delete_repo_sync(owner=repository.get('owner'), repo=repository.get('name'))
-        if remote_delete.get('error') and remote_delete.get('status_code') not in {404, None}:
-            raise HTTPException(status_code=502, detail=remote_delete.get('error'))
+        if remote_delete.get('error'):
+            manager.log_system_event(
+                component='gitea',
+                level='WARNING',
+                message=f"Remote repository delete failed for {repository.get('owner')}/{repository.get('name')}: {remote_delete.get('error')}",
+            )
 
     result = manager.delete_project(project_id)
     if result.get('status') == 'error':
         raise HTTPException(status_code=400, detail=result.get('message', 'Project deletion failed'))
     result['remote_repository_deleted'] = bool(remote_delete and not remote_delete.get('error'))
+    result['remote_repository_delete_error'] = remote_delete.get('error') if remote_delete else None
     result['remote_repository'] = repository if repository else None
     return result