- Introduction
- Getting Started with UiPath Agents
- Getting Started with UiPath Agents using LangGraph
- Building a Low-Code Agent in Studio Web
- Adding Tools to Your UiPath Agent
- Getting Started with the Flow canvas in UiPath Maestro
Introduction
Give the low-code quest classifier an API Workflow tool that looks up challenge ratings in the 5e SRD, then evaluate when and how the agent uses it.
This lab extends the quest classifier you built in Building a Low-Code Agent in Studio Web - adding an API workflow as a tool for the agent. The tool enables the agent to look up creature data from the 5e SRD (System Reference Document), providing real challenge rating data that the agent can use and cite in its reasoning.
The starting agent sorts a quest into one of four difficulty tiers, but its reasoning comes entirely from the model's training: ask it about an ancient red dragon and it answers Legendary because dragons are dangerous. With the tool added, the agent decides at runtime whether a quest needs a lookup: a quest that names a creature gets one, and a quest to deliver a letter does not.
In this lab, you:
- Build a Monster Query API workflow that searches the 5e SRD.
- Connect it to your quest classifier as a tool.
- Replace the system prompt so the agent knows when to call the tool and how to use what it returns.
- Test the agent on quests that do and do not need a lookup.
- Evaluate how the agent uses the tool with a tool-aware evaluation set.
This lab uses an API workflow as the tool. An agent can also call:
- RPA workflows
- Integration Service connectors, for systems such as Slack, Outlook, or an enterprise resource planning (ERP) system
- Document understanding
- Maestro cases and flows
- DeepRAG
- Model Context Protocol (MCP) servers
- Other agents, including remote agent-to-agent calls
Estimated time: 40-50 minutes
What you are building
| Component | Details |
|---|---|
| API Workflow | Accepts searchName (string), calls the Open5e 5e SRD API, returns a monsterResults array |
| Tool: Monster Query | The API Workflow, connected to the agent. Called only when a quest names or strongly implies a specific creature |
Agent input: questDescription | string; unchanged from the low-code lab |
Agent output: tier | string; unchanged: exactly one of Trivial, Standard, Heroic, or Legendary. When the tool returns a challenge rating, the tier follows from it |
Agent output: reasoning | string; unchanged, but now cites the challenge rating whenever the tool returned one |
| Evaluation set | A tool-aware set, scored by both a semantic similarity evaluator and a trajectory evaluator |
Prerequisites
- Studio Web - runs in your browser; no desktop installation required. Open Studio Web in Chrome, Edge, or Firefox before starting.
- UiPath account - sign up or log in to UiPath Automation Cloud before starting.
- A quest classifier agent in Studio Web - this lab builds on the agent from Building a Low-Code Agent in Studio Web, including its evaluation set. If you completed that lab, your agent is already in your workspace, and you need nothing else: the rest of this lab runs entirely in the browser.
If you need the starter agent
If you are starting fresh and did not complete the earlier lab, you can deploy the quest classifier agent and start there. The starter includes the agent and the evaluation set from the prior lab.
To restore the starter sample, you need the UiPath command-line interface (CLI). To install the CLI, refer to Environment set up and complete Step 1 (Install the UiPath CLI and Tools) and Step 4 (Authenticate to UiPath); you do not need the UiPath skills or the VS Code extension steps to complete this lab.
Use these CLI commands to deploy the starter:
git clone https://github.com/cliff-simpkins/UiPath-Workshops.git
cd UiPath-Workshops/starters/advguild.quest-classifier
uip login
uip solution upload .
git clone https://github.com/cliff-simpkins/UiPath-Workshops.git
cd UiPath-Workshops/starters/advguild.quest-classifier
uip login
uip solution upload .
Open Studio Web and confirm AdvGuild.quest-classifier appears in your workspace before beginning Step 1.