diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml
new file mode 100644
index 0000000..656ebed
--- /dev/null
+++ b/.github/workflows/ci.yml
@@ -0,0 +1,45 @@
+name: CI
+
+on:
+  push:
+    branches: [main]
+  pull_request:
+
+jobs:
+  test:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+
+      - uses: actions/setup-python@v5
+        with:
+          python-version: "3.13"
+
+      - name: Install uv
+        uses: astral-sh/setup-uv@v4
+
+      - name: Install dependencies
+        run: uv pip install --system -r requirements.txt
+
+      - name: Run unit tests
+        run: python -m pytest tests/test_db.py tests/test_repositories.py tests/test_migration.py tests/test_permissions.py tests/test_api.py tests/test_api_scripts.py tests/test_cli.py -v
+        env:
+          JWT_SECRET_KEY: ci-test-secret-32chars-minimum!!
+
+  docker-build:
+    runs-on: ubuntu-latest
+    needs: test
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: Build Docker image
+        run: docker build -t data-analyst:test .
+
+      - name: Run Docker health check
+        run: |
+          docker run -d --name test-app -p 8000:8000 \
+            -e DATA_DIR=/data -e JWT_SECRET_KEY=ci-test-secret-32chars-minimum!! \
+            data-analyst:test
+          sleep 5
+          curl -f http://localhost:8000/api/health
+          docker stop test-app
diff --git a/app/api/scripts.py b/app/api/scripts.py
new file mode 100644
index 0000000..f888c8b
--- /dev/null
+++ b/app/api/scripts.py
@@ -0,0 +1,161 @@
+"""Script management and execution endpoints."""
+
+import os
+import subprocess
+import tempfile
+import uuid
+from pathlib import Path
+
+from fastapi import APIRouter, Depends, HTTPException
+from pydantic import BaseModel
+from typing import Optional, List
+
+import duckdb
+
+from app.auth.dependencies import get_current_user, require_role, Role, _get_db
+from src.repositories.notifications import ScriptRepository
+
+router = APIRouter(prefix="/api/scripts", tags=["scripts"])
+
+SCRIPT_TIMEOUT = int(os.environ.get("SCRIPT_TIMEOUT", "300"))  # 5 min default
+SCRIPT_MAX_OUTPUT = int(os.environ.get("SCRIPT_MAX_OUTPUT", "65536"))  # 64KB
+
+
+class DeployScriptRequest(BaseModel):
+    name: str
+    source: str
+    schedule: Optional[str] = None
+
+
+class RunScriptRequest(BaseModel):
+    name: Optional[str] = None
+    source: Optional[str] = None
+
+
+class ScriptResponse(BaseModel):
+    id: str
+    name: str
+    schedule: Optional[str]
+    owner: Optional[str]
+
+
+@router.get("")
+async def list_scripts(
+    user: dict = Depends(get_current_user),
+    conn: duckdb.DuckDBPyConnection = Depends(_get_db),
+):
+    repo = ScriptRepository(conn)
+    scripts = repo.list_all()
+    return {"scripts": scripts, "count": len(scripts)}
+
+
+@router.post("/deploy", status_code=201)
+async def deploy_script(
+    request: DeployScriptRequest,
+    user: dict = Depends(get_current_user),
+    conn: duckdb.DuckDBPyConnection = Depends(_get_db),
+):
+    """Deploy a Python script to be run on the server (optionally on schedule)."""
+    repo = ScriptRepository(conn)
+    script_id = str(uuid.uuid4())
+    repo.deploy(
+        id=script_id,
+        name=request.name,
+        owner=user["id"],
+        schedule=request.schedule,
+        source=request.source,
+    )
+    return ScriptResponse(
+        id=script_id, name=request.name,
+        schedule=request.schedule, owner=user["id"],
+    )
+
+
+@router.post("/{script_id}/run")
+async def run_deployed_script(
+    script_id: str,
+    user: dict = Depends(get_current_user),
+    conn: duckdb.DuckDBPyConnection = Depends(_get_db),
+):
+    """Run a deployed script by ID."""
+    repo = ScriptRepository(conn)
+    script = repo.get(script_id)
+    if not script:
+        raise HTTPException(status_code=404, detail="Script not found")
+    return _execute_script(script["source"], script["name"])
+
+
+@router.post("/run")
+async def run_adhoc_script(
+    request: RunScriptRequest,
+    user: dict = Depends(get_current_user),
+):
+    """Run an ad-hoc Python script (not deployed)."""
+    if not request.source:
+        raise HTTPException(status_code=400, detail="Script source required")
+    return _execute_script(request.source, request.name or "adhoc")
+
+
+@router.delete("/{script_id}", status_code=204)
+async def undeploy_script(
+    script_id: str,
+    user: dict = Depends(get_current_user),
+    conn: duckdb.DuckDBPyConnection = Depends(_get_db),
+):
+    repo = ScriptRepository(conn)
+    if not repo.get(script_id):
+        raise HTTPException(status_code=404, detail="Script not found")
+    repo.undeploy(script_id)
+
+
+def _execute_script(source: str, name: str) -> dict:
+    """Execute a Python script in a sandboxed subprocess."""
+    # Safety checks
+    dangerous_imports = ["subprocess", "shutil", "ctypes", "importlib"]
+    for imp in dangerous_imports:
+        if f"import {imp}" in source or f"from {imp}" in source:
+            raise HTTPException(
+                status_code=400,
+                detail=f"Script contains disallowed import: {imp}",
+            )
+
+    data_dir = os.environ.get("DATA_DIR", "./data")
+
+    with tempfile.NamedTemporaryFile(mode="w", suffix=".py", delete=False) as f:
+        f.write(source)
+        f.flush()
+        script_path = f.name
+
+    try:
+        result = subprocess.run(
+            ["python", script_path],
+            capture_output=True,
+            text=True,
+            timeout=SCRIPT_TIMEOUT,
+            env={
+                "PATH": os.environ.get("PATH", ""),
+                "DATA_DIR": data_dir,
+                "PYTHONPATH": os.getcwd(),
+                "HOME": "/tmp",
+            },
+            cwd=os.getcwd(),
+        )
+        stdout = result.stdout[:SCRIPT_MAX_OUTPUT]
+        stderr = result.stderr[:SCRIPT_MAX_OUTPUT]
+        return {
+            "name": name,
+            "exit_code": result.returncode,
+            "stdout": stdout,
+            "stderr": stderr,
+            "truncated": len(result.stdout) > SCRIPT_MAX_OUTPUT or len(result.stderr) > SCRIPT_MAX_OUTPUT,
+        }
+    except subprocess.TimeoutExpired:
+        return {
+            "name": name,
+            "exit_code": -1,
+            "stdout": "",
+            "stderr": f"Script timed out after {SCRIPT_TIMEOUT}s",
+            "truncated": False,
+        }
+    finally:
+        os.unlink(script_path)
diff --git a/app/api/settings.py b/app/api/settings.py
new file mode 100644
index 0000000..60fa34c
--- /dev/null
+++ b/app/api/settings.py
@@ -0,0 +1,53 @@
+"""User sync settings endpoints."""
+
+from fastapi import APIRouter, Depends, HTTPException
+from pydantic import BaseModel
+from typing import Optional, List
+
+import duckdb
+
+from app.auth.dependencies import get_current_user, _get_db
+from src.repositories.sync_settings import SyncSettingsRepository, DatasetPermissionRepository
+
+router = APIRouter(prefix="/api/settings", tags=["settings"])
+
+
+class DatasetSettingRequest(BaseModel):
+    dataset: str
+    enabled: bool
+
+
+@router.get("")
+async def get_settings(
+    user: dict = Depends(get_current_user),
+    conn: duckdb.DuckDBPyConnection = Depends(_get_db),
+):
+    """Get current user's sync settings and permissions."""
+    settings_repo = SyncSettingsRepository(conn)
+    perm_repo = DatasetPermissionRepository(conn)
+
+    settings = settings_repo.get_user_settings(user["id"])
+    permissions = perm_repo.get_user_permissions(user["id"])
+
+    return {
+        "user_id": user["id"],
+        "sync_settings": settings,
+        "permissions": permissions,
+    }
+
+
+@router.put("/dataset")
+async def update_dataset_setting(
+    request: DatasetSettingRequest,
+    user: dict = Depends(get_current_user),
+    conn: duckdb.DuckDBPyConnection = Depends(_get_db),
+):
+    """Enable or disable a dataset for sync."""
+    # Check permission
+    perm_repo = DatasetPermissionRepository(conn)
+    if not perm_repo.has_access(user["id"], request.dataset):
+        raise HTTPException(status_code=403, detail=f"No access to dataset '{request.dataset}'")
+
+    settings_repo = SyncSettingsRepository(conn)
+    settings_repo.set_dataset_enabled(user["id"], request.dataset, request.enabled)
+    return {"dataset": request.dataset, "enabled": request.enabled}
diff --git a/app/main.py b/app/main.py
index 4456db9..0c10d87 100644
--- a/app/main.py
+++ b/app/main.py
@@ -11,6 +11,8 @@ from app.api.query import router as query_router
 from app.api.users import router as users_router
 from app.api.memory import router as memory_router
 from app.api.upload import router as upload_router
+from app.api.scripts import router as scripts_router
+from app.api.settings import router as settings_router
 
 
 def create_app() -> FastAPI:
@@ -38,6 +40,8 @@ def create_app() -> FastAPI:
     app.include_router(users_router)
     app.include_router(memory_router)
     app.include_router(upload_router)
+    app.include_router(scripts_router)
+    app.include_router(settings_router)
 
     return app
 
diff --git a/cli/skills/connectors.md b/cli/skills/connectors.md
new file mode 100644
index 0000000..3b451a2
--- /dev/null
+++ b/cli/skills/connectors.md
@@ -0,0 +1,30 @@
+# Connectors — How to add a new data source
+
+## Existing Connectors
+- **Keboola** (`connectors/keboola/`) — Keboola Storage API
+- **BigQuery** (`connectors/bigquery/`) — Google BigQuery
+- **Jira** (`connectors/jira/`) — Jira webhook + API
+
+## Adding a New Connector
+
+1. Create `connectors/<name>/adapter.py` implementing the `DataSource` ABC:
+   ```python
+   from src.data_sync import DataSource
+
+   class MyDataSource(DataSource):
+       def sync_table(self, table_config, sync_state): ...
+       def discover_tables(self): ...
+       def get_column_metadata(self, table_id): ...
+       def get_source_name(self): ...
+   ```
+
+2. The factory in `src/data_sync.py:create_data_source()` auto-discovers connectors.
+   Set `DATA_SOURCE=<name>` in instance.yaml or .env.
+
+3. Add required env vars to `.env` and `config/.env.template`.
+
+4. Add tests to `tests/test_<name>_adapter.py`.
+
+## Configuration
+Each connector reads credentials from environment variables.
+Table definitions are in `docs/data_description.md` (YAML blocks).
diff --git a/cli/skills/corporate-memory.md b/cli/skills/corporate-memory.md
new file mode 100644
index 0000000..e91329d
--- /dev/null
+++ b/cli/skills/corporate-memory.md
@@ -0,0 +1,30 @@
+# Corporate Memory — Knowledge sharing and governance
+
+## What It Is
+Corporate memory collects knowledge from all analysts' CLAUDE.local.md files
+and makes it available to everyone through a curated catalog.
+
+## How It Works
+1. Analysts write insights in their CLAUDE.local.md
+2. `da sync --upload-only` pushes content to server
+3. Server processes with LLM (Haiku) to extract knowledge items
+4. Items go through governance (pending → approved/mandatory)
+5. Approved items are distributed as Claude rules
+
+## Governance Flow
+- **pending**: New item, awaiting review
+- **approved**: Available to all users
+- **mandatory**: Force-pushed to all users' rules
+- **rejected**: Not distributed
+
+## Admin Commands
+```bash
+# View pending items (via web UI or API)
+da query "SELECT id, title, status FROM system.knowledge_items WHERE status='pending'" --remote
+
+# Approve/reject via API
+curl -X PUT http://server:8000/api/memory/<id>/status?new_status=approved -H "Authorization: Bearer $TOKEN"
+```
+
+## Voting
+Users can upvote/downvote knowledge items to surface the most useful ones.
diff --git a/cli/skills/notifications.md b/cli/skills/notifications.md
new file mode 100644
index 0000000..1568128
--- /dev/null
+++ b/cli/skills/notifications.md
@@ -0,0 +1,37 @@
+# Notifications — How notifications work
+
+## Architecture
+1. User creates a Python script (locally or via Claude Code)
+2. Script queries local DuckDB and produces output
+3. Output is sent via Telegram bot or WebSocket gateway
+
+## Creating a Notification Script
+```python
+# user/scripts/sales_alert.py
+"""Sales alert - checks daily revenue."""
+import duckdb
+
+conn = duckdb.connect('user/duckdb/analytics.duckdb', read_only=True)
+result = conn.execute("SELECT sum(amount) as revenue FROM orders WHERE date = current_date").fetchone()
+print(f"Today's revenue: ${result[0]:,.2f}")
+```
+
+## Running Locally
+```bash
+da scripts run sales_alert           # runs on your machine
+```
+
+## Deploying to Server
+```bash
+da scripts deploy sales_alert --schedule "0 8 * * MON"  # every Monday 8 AM
+```
+
+## Delivery Channels
+- **Telegram**: Link via `da auth telegram-link`
+- **Desktop app**: Via WebSocket gateway (automatic if connected)
+
+## Managing Scripts
+```bash
+da scripts list                      # all deployed scripts
+da scripts undeploy <script-id>      # remove from server
+```
diff --git a/cli/skills/security.md b/cli/skills/security.md
new file mode 100644
index 0000000..9a36e5a
--- /dev/null
+++ b/cli/skills/security.md
@@ -0,0 +1,37 @@
+# Security — RBAC, permissions, and audit
+
+## Roles
+| Role | Permissions |
+|------|-------------|
+| `viewer` | Read catalog, view profiles, browse corporate memory |
+| `analyst` | + sync data, run queries, vote, run/deploy scripts |
+| `admin` | + manage users, approve knowledge, trigger sync |
+| `km_admin` | + corporate memory governance |
+
+## Managing Users
+```bash
+da admin add-user user@company.com --role analyst
+da admin list-users
+da admin remove-user <user-id>
+```
+
+## Dataset Permissions
+Admins grant dataset access per user. Users can only sync datasets they have access to.
+
+## Audit Trail
+Every API call is logged. Query with:
+```bash
+da query "SELECT * FROM system.audit_log ORDER BY timestamp DESC LIMIT 20" --remote
+```
+
+## Script Sandboxing
+User scripts run in isolated subprocess with:
+- Limited environment (no access to secrets)
+- Timeout (default 5 min)
+- Blocked imports (subprocess, shutil, ctypes)
+- Stdout/stderr size cap (64KB)
+
+## JWT Tokens
+- Issued on login, valid 30 days
+- Contains: user_id, email, role
+- Set JWT_SECRET_KEY in .env (min 32 chars)
diff --git a/config/deploy.staging.yml b/config/deploy.staging.yml
new file mode 100644
index 0000000..9d74e07
--- /dev/null
+++ b/config/deploy.staging.yml
@@ -0,0 +1,20 @@
+# Kamal staging deployment config
+# Usage: kamal deploy -d staging
+
+servers:
+  web:
+    hosts:
+      - YOUR_STAGING_SERVER_IP
+
+accessories:
+  scheduler:
+    host: YOUR_STAGING_SERVER_IP
+  telegram-bot:
+    host: YOUR_STAGING_SERVER_IP
+
+proxy:
+  host: staging.data.your-domain.com
+
+env:
+  clear:
+    LOG_LEVEL: debug
diff --git a/config/deploy.yml b/config/deploy.yml
new file mode 100644
index 0000000..444c8b6
--- /dev/null
+++ b/config/deploy.yml
@@ -0,0 +1,67 @@
+# Kamal production deployment config
+# Usage: kamal deploy
+
+service: data-analyst
+
+image: ghcr.io/keboola/data-analyst
+
+registry:
+  server: ghcr.io
+  username:
+    - KAMAL_REGISTRY_USERNAME
+  password:
+    - KAMAL_REGISTRY_PASSWORD
+
+servers:
+  web:
+    hosts:
+      - YOUR_SERVER_IP
+    cmd: uvicorn app.main:app --host 0.0.0.0 --port 8000
+    options:
+      volume:
+        - /data:/data
+
+accessories:
+  scheduler:
+    image: ghcr.io/keboola/data-analyst
+    host: YOUR_SERVER_IP
+    cmd: python -m services.scheduler
+    env:
+      clear:
+        API_URL: http://data-analyst-web:8000
+      secret:
+        - SCHEDULER_API_TOKEN
+    volumes:
+      - /data:/data
+
+  telegram-bot:
+    image: ghcr.io/keboola/data-analyst
+    host: YOUR_SERVER_IP
+    cmd: python -m services.telegram_bot
+    env:
+      secret:
+        - TELEGRAM_BOT_TOKEN
+    volumes:
+      - /data:/data
+
+proxy:
+  ssl: true
+  host: data.your-domain.com
+
+healthcheck:
+  path: /api/health
+  port: 8000
+  interval: 30
+
+env:
+  clear:
+    DATA_DIR: /data
+    LOG_LEVEL: info
+  secret:
+    - JWT_SECRET_KEY
+    - KEBOOLA_STORAGE_TOKEN
+    - KEBOOLA_STACK_URL
+    - KEBOOLA_PROJECT_ID
+    - GOOGLE_CLIENT_ID
+    - GOOGLE_CLIENT_SECRET
+    - TELEGRAM_BOT_TOKEN
diff --git a/src/repositories/sync_settings.py b/src/repositories/sync_settings.py
new file mode 100644
index 0000000..2d45c30
--- /dev/null
+++ b/src/repositories/sync_settings.py
@@ -0,0 +1,87 @@
+"""Repository for user sync settings and dataset permissions."""
+
+from datetime import datetime, timezone
+from typing import Any, Optional, List, Dict
+
+import duckdb
+
+
+class SyncSettingsRepository:
+    def __init__(self, conn: duckdb.DuckDBPyConnection):
+        self.conn = conn
+
+    def get_user_settings(self, user_id: str) -> List[Dict[str, Any]]:
+        results = self.conn.execute(
+            "SELECT * FROM user_sync_settings WHERE user_id = ? ORDER BY dataset",
+            [user_id],
+        ).fetchall()
+        if not results:
+            return []
+        columns = [desc[0] for desc in self.conn.description]
+        return [dict(zip(columns, row)) for row in results]
+
+    def set_dataset_enabled(self, user_id: str, dataset: str, enabled: bool) -> None:
+        now = datetime.now(timezone.utc)
+        self.conn.execute(
+            """INSERT INTO user_sync_settings (user_id, dataset, enabled, updated_at)
+            VALUES (?, ?, ?, ?)
+            ON CONFLICT (user_id, dataset) DO UPDATE SET enabled = excluded.enabled, updated_at = excluded.updated_at""",
+            [user_id, dataset, enabled, now],
+        )
+
+    def is_dataset_enabled(self, user_id: str, dataset: str) -> bool:
+        result = self.conn.execute(
+            "SELECT enabled FROM user_sync_settings WHERE user_id = ? AND dataset = ?",
+            [user_id, dataset],
+        ).fetchone()
+        return bool(result and result[0])
+
+    def get_enabled_datasets(self, user_id: str) -> List[str]:
+        results = self.conn.execute(
+            "SELECT dataset FROM user_sync_settings WHERE user_id = ? AND enabled = true",
+            [user_id],
+        ).fetchall()
+        return [r[0] for r in results]
+
+
+class DatasetPermissionRepository:
+    def __init__(self, conn: duckdb.DuckDBPyConnection):
+        self.conn = conn
+
+    def grant(self, user_id: str, dataset: str, access: str = "read") -> None:
+        self.conn.execute(
+            """INSERT INTO dataset_permissions (user_id, dataset, access)
+            VALUES (?, ?, ?)
+            ON CONFLICT (user_id, dataset) DO UPDATE SET access = excluded.access""",
+            [user_id, dataset, access],
+        )
+
+    def revoke(self, user_id: str, dataset: str) -> None:
+        self.conn.execute(
+            "DELETE FROM dataset_permissions WHERE user_id = ? AND dataset = ?",
+            [user_id, dataset],
+        )
+
+    def has_access(self, user_id: str, dataset: str) -> bool:
+        result = self.conn.execute(
+            "SELECT access FROM dataset_permissions WHERE user_id = ? AND dataset = ?",
+            [user_id, dataset],
+        ).fetchone()
+        return result is not None and result[0] != "none"
+
+    def get_user_permissions(self, user_id: str) -> List[Dict[str, Any]]:
+        results = self.conn.execute(
+            "SELECT * FROM dataset_permissions WHERE user_id = ? ORDER BY dataset",
+            [user_id],
+        ).fetchall()
+        if not results:
+            return []
+        columns = [desc[0] for desc in self.conn.description]
+        return [dict(zip(columns, row)) for row in results]
+
+    def get_accessible_datasets(self, user_id: str) -> List[str]:
+        results = self.conn.execute(
+            "SELECT dataset FROM dataset_permissions WHERE user_id = ? AND access != 'none'",
+            [user_id],
+        ).fetchall()
+        return [r[0] for r in results]
diff --git a/tests/test_api_scripts.py b/tests/test_api_scripts.py
new file mode 100644
index 0000000..b9dcae7
--- /dev/null
+++ b/tests/test_api_scripts.py
@@ -0,0 +1,126 @@
+"""Tests for scripts and settings API endpoints."""
+
+import os
+import pytest
+from fastapi.testclient import TestClient
+
+
+@pytest.fixture
+def client(tmp_path):
+    os.environ["DATA_DIR"] = str(tmp_path)
+    os.environ["JWT_SECRET_KEY"] = "test-secret-32chars-minimum!!!!!"
+    os.environ["SCRIPT_TIMEOUT"] = "10"
+
+    from app.main import create_app
+    from src.db import get_system_db
+    from src.repositories.users import UserRepository
+    from src.repositories.sync_settings import DatasetPermissionRepository
+    from app.auth.jwt import create_access_token
+
+    conn = get_system_db()
+    user_repo = UserRepository(conn)
+    user_repo.create(id="admin1", email="admin@acme.com", name="Admin", role="admin")
+    user_repo.create(id="analyst1", email="analyst@acme.com", name="Analyst", role="analyst")
+
+    perm_repo = DatasetPermissionRepository(conn)
+    perm_repo.grant("analyst1", "sales", "read")
+    perm_repo.grant("analyst1", "support", "read")
+    conn.close()
+
+    app = create_app()
+    test_client = TestClient(app)
+    admin_token = create_access_token("admin1", "admin@acme.com", "admin")
+    analyst_token = create_access_token("analyst1", "analyst@acme.com", "analyst")
+
+    return test_client, admin_token, analyst_token
+
+
+class TestScriptsAPI:
+    def test_list_scripts_empty(self, client):
+        c, _, analyst_token = client
+        resp = c.get("/api/scripts", headers={"Authorization": f"Bearer {analyst_token}"})
+        assert resp.status_code == 200
+        assert resp.json()["count"] == 0
+
+    def test_deploy_and_list(self, client):
+        c, _, analyst_token = client
+        headers = {"Authorization": f"Bearer {analyst_token}"}
+
+        resp = c.post("/api/scripts/deploy", json={
+            "name": "hello", "source": "print('hello world')",
+        }, headers=headers)
+        assert resp.status_code == 201
+        script_id = resp.json()["id"]
+
+        resp = c.get("/api/scripts", headers=headers)
+        assert resp.json()["count"] == 1
+
+    def test_run_script(self, client):
+        c, _, analyst_token = client
+        headers = {"Authorization": f"Bearer {analyst_token}"}
+
+        resp = c.post("/api/scripts/run", json={
+            "source": "print('hello from script')", "name": "test",
+        }, headers=headers)
+        assert resp.status_code == 200
+        data = resp.json()
+        assert data["exit_code"] == 0
+        assert "hello from script" in data["stdout"]
+
+    def test_run_blocked_import(self, client):
+        c, _, analyst_token = client
+        headers = {"Authorization": f"Bearer {analyst_token}"}
+
+        resp = c.post("/api/scripts/run", json={
+            "source": "import subprocess; subprocess.run(['ls'])", "name": "bad",
+        }, headers=headers)
+        assert resp.status_code == 400
+        assert "disallowed" in resp.json()["detail"]
+
+    def test_deploy_run_undeploy(self, client):
+        c, _, analyst_token = client
+        headers = {"Authorization": f"Bearer {analyst_token}"}
+
+        # Deploy
+        resp = c.post("/api/scripts/deploy", json={
+            "name": "calc", "source": "print(2+2)", "schedule": "0 8 * * MON",
+        }, headers=headers)
+        script_id = resp.json()["id"]
+
+        # Run
+        resp = c.post(f"/api/scripts/{script_id}/run", headers=headers)
+        assert resp.status_code == 200
+        assert "4" in resp.json()["stdout"]
+
+        # Undeploy
+        resp = c.delete(f"/api/scripts/{script_id}", headers=headers)
+        assert resp.status_code == 204
+
+
+class TestSettingsAPI:
+    def test_get_settings(self, client):
+        c, _, analyst_token = client
+        resp = c.get("/api/settings", headers={"Authorization": f"Bearer {analyst_token}"})
+        assert resp.status_code == 200
+        data = resp.json()
+        assert data["user_id"] == "analyst1"
+        assert len(data["permissions"]) == 2
+
+    def test_enable_dataset(self, client):
+        c, _, analyst_token = client
+        headers = {"Authorization": f"Bearer {analyst_token}"}
+
+        resp = c.put("/api/settings/dataset", json={
+            "dataset": "sales", "enabled": True,
+        }, headers=headers)
+        assert resp.status_code == 200
+        assert resp.json()["enabled"] is True
+
+    def test_enable_unauthorized_dataset(self, client):
+        c, _, analyst_token = client
+        headers = {"Authorization": f"Bearer {analyst_token}"}
+
+        resp = c.put("/api/settings/dataset", json={
+            "dataset": "hr_secret", "enabled": True,
+        }, headers=headers)
+        assert resp.status_code == 403
diff --git a/tests/test_permissions.py b/tests/test_permissions.py
new file mode 100644
index 0000000..976d777
--- /dev/null
+++ b/tests/test_permissions.py
@@ -0,0 +1,80 @@
+"""Tests for sync settings, dataset permissions, and script execution."""
+
+import os
+import pytest
+
+
+@pytest.fixture
+def db_conn(tmp_path):
+    os.environ["DATA_DIR"] = str(tmp_path)
+    from src.db import get_system_db
+    conn = get_system_db()
+    yield conn
+    conn.close()
+
+
+class TestSyncSettingsRepository:
+    def test_set_and_get(self, db_conn):
+        from src.repositories.sync_settings import SyncSettingsRepository
+        repo = SyncSettingsRepository(db_conn)
+        repo.set_dataset_enabled("u1", "sales", True)
+        repo.set_dataset_enabled("u1", "support", False)
+        settings = repo.get_user_settings("u1")
+        assert len(settings) == 2
+
+    def test_is_enabled(self, db_conn):
+        from src.repositories.sync_settings import SyncSettingsRepository
+        repo = SyncSettingsRepository(db_conn)
+        repo.set_dataset_enabled("u1", "sales", True)
+        assert repo.is_dataset_enabled("u1", "sales") is True
+        assert repo.is_dataset_enabled("u1", "support") is False
+
+    def test_get_enabled_datasets(self, db_conn):
+        from src.repositories.sync_settings import SyncSettingsRepository
+        repo = SyncSettingsRepository(db_conn)
+        repo.set_dataset_enabled("u1", "sales", True)
+        repo.set_dataset_enabled("u1", "support", False)
+        repo.set_dataset_enabled("u1", "hr", True)
+        enabled = repo.get_enabled_datasets("u1")
+        assert set(enabled) == {"sales", "hr"}
+
+    def test_toggle_dataset(self, db_conn):
+        from src.repositories.sync_settings import SyncSettingsRepository
+        repo = SyncSettingsRepository(db_conn)
+        repo.set_dataset_enabled("u1", "sales", True)
+        assert repo.is_dataset_enabled("u1", "sales") is True
+        repo.set_dataset_enabled("u1", "sales", False)
+        assert repo.is_dataset_enabled("u1", "sales") is False
+
+
+class TestDatasetPermissionRepository:
+    def test_grant_and_check(self, db_conn):
+        from src.repositories.sync_settings import DatasetPermissionRepository
+        repo = DatasetPermissionRepository(db_conn)
+        repo.grant("u1", "sales", "read")
+        assert repo.has_access("u1", "sales") is True
+        assert repo.has_access("u1", "hr") is False
+
+    def test_revoke(self, db_conn):
+        from src.repositories.sync_settings import DatasetPermissionRepository
+        repo = DatasetPermissionRepository(db_conn)
+        repo.grant("u1", "sales", "read")
+        repo.revoke("u1", "sales")
+        assert repo.has_access("u1", "sales") is False
+
+    def test_get_accessible_datasets(self, db_conn):
+        from src.repositories.sync_settings import DatasetPermissionRepository
+        repo = DatasetPermissionRepository(db_conn)
+        repo.grant("u1", "sales", "read")
+        repo.grant("u1", "hr", "read")
+        repo.grant("u1", "finance", "none")
+        accessible = repo.get_accessible_datasets("u1")
+        assert set(accessible) == {"sales", "hr"}
+
+    def test_get_user_permissions(self, db_conn):
+        from src.repositories.sync_settings import DatasetPermissionRepository
+        repo = DatasetPermissionRepository(db_conn)
+        repo.grant("u1", "sales", "read")
+        repo.grant("u1", "hr", "read")
+        perms = repo.get_user_permissions("u1")
+        assert len(perms) == 2