agents
latest
false
- 基本情報
- Studio Web での UiPath Agents
- UiPath のコード化されたエージェント
重要 :
このコンテンツの一部は機械翻訳によって処理されており、完全な翻訳を保証するものではありません。
新しいコンテンツの翻訳は、およそ 1 ~ 2 週間で公開されます。

Agents ガイド
最終更新日時 2026年4月6日
Autopilot for Agents tools
Autopilot for Agents uses Framework Tools to read and modify your agent project in response to prompts. Each tool targets a specific part of the agent definition or platform resources. All tools are enabled by default and can be individually toggled in Settings > Tools > Framework Tools.
The tables below describe each tool and what it acts on.
Agent definition tools
These tools read and update components of the agent definition directly.
| ツール | 説明 |
|---|---|
| エージェントの設定 | Provides Autopilot with the current agent configuration as context for generating suggestions. |
| システム プロンプト | Reads and updates the agent's system prompt. |
| ユーザー プロンプト | Reads and updates the agent's user prompt. |
| 入力スキーマ | Reads and updates the agent's input schema. |
| Output schema | Reads and updates the agent's output schema. |
| Tool name | Reads and updates the name of a configured agent tool. |
| Tool description | Reads and updates the description of a configured agent tool. |
| Agent tool | Adds an RPA workflow, API workflow, or other process as a tool available to the agent. |
| Built-in tool | Manages UiPath built-in tools for the agent. |
Resource and evaluation tools
These tools search for external resources and support evaluation setup.
| ツール | 説明 |
|---|---|
| Get Solution Resources | Searches for available solution resources to add to the agent, including RPA workflows, API workflows, agent processes, and Context Grounding (ECS) context sources. |
| Agent context | Adds a Context Grounding context source to the agent to support knowledge retrieval during execution. |
| Search activities | Searches for available Integration Service activities by keyword or connector key. |
| Integration service activity | Validates and adds an Integration Service activity as a tool for the agent. |
| Create evaluation set | Creates an evaluation set for testing and scoring agent responses. |
注:
Tools marked as experimental in the Settings panel may change behavior or availability in future releases.
関連トピック
- Autopilot for Agents — overview, interaction modes, settings, and LLM options