chore: add agent using API resume trigger

Author: PopescuTudor

samples/expense-approval-agent/.env.example

@@ -0,0 +1,2 @@
+UIPATH_URL=https://alpha.uipath.com/ada/byoa
+UIPATH_ACCESS_TOKEN=xxx

samples/expense-approval-agent/README.md

@@ -0,0 +1,173 @@
+# Simplified Expense Approval Agent
+
+A streamlined expense approval workflow that validates expenses and always suspends for manager review via API trigger.
+
+## Features
+
+- **Basic Validation**: Checks receipt, date, and amount limits
+- **Always Suspends**: Every expense goes to manager for review
+- **API Trigger Resume**: Manager approves/rejects via API
+- **Clear Audit Trail**: Processing notes track all decisions
+- **No LLM Required**: Pure business logic
+
+## How It Works
+
+### Workflow Steps
+
+1. **Prepare Input**: Generates unique expense ID (EXP-YYYYMMDD-XXXXXXXX)
+2. **Validate Expense**: Simple checks for receipt, date, and limits
+3. **Manager Review**: Always suspends for approval via API trigger
+4. **Finalize**: Returns final decision with notes
+
+### Validation Rules
+
+- **Receipt Required**: For expenses over $25
+- **Date Limit**: Must be within 90 days
+- **Category Limits**:
+  - Travel: $2000
+  - Meals: $150
+  - Supplies: $500
+  - Training: $5000
+  - Equipment: $3000
+  - Other: $300
+
+## Setup
+
+### 1. Install Dependencies
+
+```bash
+cd expense-approval-agent
+pip install uv
+uv venv -p 3.11 .venv
+.venv\Scripts\activate  # Windows
+source .venv/bin/activate  # Linux/Mac
+uv sync
+```
+
+### 2. Configure Environment
+
+Create `.env` file:
+```env
+UIPATH_URL=https://your-orchestrator.com
+UIPATH_ACCESS_TOKEN=your_token
+```
+
+### 3. Test Locally
+
+```bash
+uipath run agent '{
+  "employee_id": "EMP001",
+  "employee_name": "Jane Smith",
+  "department": "sales",
+  "expense_amount": 450.00,
+  "expense_category": "travel",
+  "expense_description": "Client meeting in Boston - flight and hotel",
+  "receipt_attached": true,
+  "expense_date": "2024-01-15",
+  "manager_email": "manager@company.com"
+}'
+```
+
+## Usage Examples
+
+### Example 1: Valid Expense
+```json
+{
+  "employee_id": "EMP002",
+  "employee_name": "John Doe",
+  "department": "engineering",
+  "expense_amount": 45.50,
+  "expense_category": "meals",
+  "expense_description": "Team lunch during sprint planning",
+  "receipt_attached": true,
+  "expense_date": "2025-09-20"
+}
+```
+Result: Validates successfully, suspends for manager review
+
+### Example 2: Expense with Warnings
+```json
+{
+  "employee_id": "EMP003",
+  "employee_name": "Alice Johnson",
+  "department": "hr",
+  "expense_amount": 2500.00,
+  "expense_category": "travel",
+  "expense_description": "Flight and hotel for conference",
+  "receipt_attached": true,
+  "expense_date": "2025-09-10"
+}
+```
+Result: Warning (exceeds $2000 limit), suspends for manager review
+
+## Resume Options
+
+When the agent suspends for manager review, you have two options:
+
+### Option 1: Orchestrator UI
+In the "Human review required" popup, simply type:
+- `true` to approve
+- `false` to reject
+
+### Option 2: API Call
+Resume via API with JSON payload:
+```json
+{
+  "approved": true
+}
+```
+
+or
+
+```json
+{
+  "approved": false
+}
+```
+
+## Output Format
+
+```json
+{
+  "expense_id": "EXP-20251001-ABC123",
+  "status": "approved",
+  "approval_level": "manager",
+  "final_amount": 45.5,
+  "processing_notes": [
+    "Expense report EXP-20251001-ABC123 created at 2025-10-01T11:53:07.354154",
+    "Approved on 2025-10-01 11:54",
+    "Reimbursement by 2025-10-06"
+  ],
+  "reimbursement_eta": "2025-10-06"
+}
+```
+
+## Deployment
+
+### 1. Package and Publish
+
+```bash
+uipath auth
+uipath pack
+uipath publish
+```
+
+### 2. Monitor in Orchestrator
+
+- View suspended expenses awaiting approval
+- Check processing logs and audit trails
+- Approve/reject directly from Orchestrator UI or call `/orchestrator_/api/JobTriggers/DeliverPayload/{inbox_id}` to resume
+
+## Test Cases
+
+The `test_cases/` folder contains comprehensive test scenarios with **deterministic outputs**:
+- `input/` - Test input files with various expense scenarios
+- `expected_output/` - Expected results for each test case
+
+**Static Values for Reliable Testing:**
+- Expense IDs: Generated from employee_id hash (e.g., EMP001 → EXP-20251001-B86D07B6)
+- Creation time: Always 2025-10-01T10:00:00.000000
+- Approval time: Always 2025-10-01 10:05
+- Reimbursement date: Always 2025-10-06 for approved expenses
+
+This ensures the same input always produces the same output, making tests reliable and repeatable.

samples/expense-approval-agent/agent.mermaid

@@ -0,0 +1,20 @@
+---
+config:
+  flowchart:
+    curve: linear
+---
+graph TD;
+	__start__([<p>__start__</p>]):::first
+	prepare_input(prepare_input)
+	validate_expense(validate_expense)
+	manager_review(manager_review)
+	finalize_expense(finalize_expense)
+	__end__([<p>__end__</p>]):::last
+	__start__ --> prepare_input;
+	manager_review --> finalize_expense;
+	prepare_input --> validate_expense;
+	validate_expense --> manager_review;
+	finalize_expense --> __end__;
+	classDef default fill:#f2f0ff,line-height:1.2
+	classDef first fill-opacity:0
+	classDef last fill:#bfb6fc

samples/expense-approval-agent/langgraph.json

@@ -0,0 +1,7 @@
+{
+    "dependencies": ["."],
+    "graphs": {
+      "agent": "./main.py:graph"
+    },
+    "env": ".env"
+  }