# AI use cases for Unreal Agent (unreallabsai/unreal-agent) Repo: [unreallabsai/unreal-agent](https://github.com/unreallabsai/unreal-agent) | Use cases: 4 | Page: https://dedupedai.com/use-cases/unreallabsai-unreal-agent/ Unreal Agent is a Go library and set of programs for building async-first AI agents, including a cloud agent runner and a terminal UI. This file holds every use case of this repo. Each one starts with a "## " heading and ends with its URL. - Run async agent tasks on the cloud runner: https://dedupedai.com/use-cases/unreallabsai-unreal-agent/run-async-agent-cloud-runner/ - Run a coding agent from the terminal TUI: https://dedupedai.com/use-cases/unreallabsai-unreal-agent/use-agent-in-terminal-tui/ - Build a custom agent tool on the harness library: https://dedupedai.com/use-cases/unreallabsai-unreal-agent/build-custom-agent-on-harness/ - Run agent tools inside a remote sandbox: https://dedupedai.com/use-cases/unreallabsai-unreal-agent/run-agent-tools-in-remote-sandbox/ ## How to run async agent tasks on the cloud runner with Unreal Agent Repo: [unreallabsai/unreal-agent](https://github.com/unreallabsai/unreal-agent) | Difficulty: Advanced | Time: 90 minutes | Basis: From the README Submit tasks to the cloud agent runner with unique input IDs so redelivered work is not repeated. **Who it is for:** A developer or platform engineer who needs an agent to process tasks in the background without a person watching each step. **The problem:** Background agent work can be delivered more than once, and repeated runs can duplicate side effects. **The outcome:** Tasks run through the cloud agent runner, and each input is recorded in the session history with a stable ID. ### Steps 1. `git clone https://github.com/unreallabsai/unreal-agent` (inferred, check the README) 2. Open cmd/unreal-agent-runner and read the source to find its entry point and how it accepts input. 3. `go build ./cmd/unreal-agent-runner` (inferred, check the README) 4. Submit your task as one input with a caller-supplied globally unique ID. 5. Resubmit the same input ID and confirm the session does not process it a second time. ### Prompt to paste into your agent ```text Using the unreal-agent repo at , build cmd/unreal-agent-runner and submit the task as one input with the unique ID . Expected output: a session history entry for that input and its result. Verify that sending the same input ID again does not run the task twice within the same session. ``` ### Caveat The inbox that drops duplicate inputs is session-scoped and in-memory, so duplicate protection does not extend across separate sessions. Risk, credentials: The LLM Adapter handles provider authentication, so keep provider keys out of task text, logs, and session history. Dual-use notice: only use this on software, sites and systems you own or are authorized to test. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/unreallabsai-unreal-agent/run-async-agent-cloud-runner/ ## How to run a coding agent from the terminal TUI with Unreal Agent Repo: [unreallabsai/unreal-agent](https://github.com/unreallabsai/unreal-agent) | Difficulty: Intermediate | Time: 45 minutes | Basis: From the README Start the cozy TUI and give the agent a local task that uses Bash or image viewing. **Who it is for:** A developer who wants to try the agent interactively in a terminal before wiring it into anything bigger. **The problem:** Setting up an agent harness before seeing it work makes it hard to judge whether it fits a task. **The outcome:** A working terminal session where the agent uses its built-in tools and the history can be reviewed. ### Steps 1. `git clone https://github.com/unreallabsai/unreal-agent` (inferred, check the README) 2. `go build ./cmd/unreal-agent-tui` (inferred, check the README) 3. Start the built TUI binary in a terminal and open a new session. 4. Give the agent a small task that needs the Bash tool or the ViewImage tool. 5. Review the session history to see the tool calls and their results. ### Prompt to paste into your agent ```text In the unreal-agent repo at , build cmd/unreal-agent-tui and start a session. Give the agent the task . Expected output: the agent's answer plus the recorded tool calls in the session. Verify that every tool call shown in the history matches an action you approved. ``` ### Caveat The README describes the TUI only as cozy, so check its keys and options in the cmd/unreal-agent-tui source before relying on them. Risk, remote script: The Bash tool can run shell commands on your machine, so start with a task in a disposable directory. Dual-use notice: only use this on software, sites and systems you own or are authorized to test. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/unreallabsai-unreal-agent/use-agent-in-terminal-tui/ ## How to build a custom agent tool on the harness library with Unreal Agent Repo: [unreallabsai/unreal-agent](https://github.com/unreallabsai/unreal-agent) | Difficulty: Advanced | Time: 120 minutes | Basis: From the README Write a tool with a schema and translator that turns model tool calls into async operations. **Who it is for:** A Go engineer building an in-house agent that needs its own tools, such as a ticket lookup or a file workflow. **The problem:** An off-the-shelf agent does not know the internal systems a team needs it to call. **The outcome:** A custom tool registered with the harness, where translators validate calls and produce operations that run asynchronously. ### Steps 1. `git clone https://github.com/unreallabsai/unreal-agent` (inferred, check the README) 2. Read the harness/ library and the glossary to learn the Tool, Tool translator, and Operation terms. 3. Define your tool with a schema and bind it to a translator that performs no I/O. 4. Make the translator return a tool call status and serializable operations, using the primitives in harness/primitives/ where they fit. 5. Run the coordinator with your tool in the host-selected set and check that the operation result is formatted into the model result. ### Prompt to paste into your agent ```text In the unreal-agent repo at , add a tool named with a schema and a translator that returns serializable operations for . Expected output: a model-facing result produced from the operation output. Verify that the translator performs no I/O and that the operation is serializable. ``` ### Caveat The README states that translators must not perform I/O or suspend the event loop, so any external lookup belongs in an operation, not in the translator. Dual-use notice: only use this on software, sites and systems you own or are authorized to test. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/unreallabsai-unreal-agent/build-custom-agent-on-harness/ ## How to run agent tools inside a remote sandbox with Unreal Agent Repo: [unreallabsai/unreal-agent](https://github.com/unreallabsai/unreal-agent) | Difficulty: Advanced | Time: 120 minutes | Basis: From the README Send serialized operations from a proxy operations manager to a local manager inside a remote sandbox. **Who it is for:** A platform engineer who wants agent tool execution isolated from the host machine and run in a remote sandbox. **The problem:** Letting an agent run commands on the host creates risk that the host cannot easily contain. **The outcome:** Tool operations execute in the remote sandbox and their results return to the agent as model results. ### Steps 1. `git clone https://github.com/unreallabsai/unreal-agent` (inferred, check the README) 2. Start a local operations manager in a process inside the remote sandbox. 3. Configure a proxy operations manager on the host to send serialized operations to that sandbox process. 4. Select the Bash tool in the host-selected tool set so its operations go through the proxy. 5. Check that the output of a sandbox command returns to the session as the tool result. ### Prompt to paste into your agent ```text Using the unreal-agent repo at , set up a proxy operations manager that sends serialized operations to a local operations manager in . Expected output: the result of a harmless Bash command run inside the sandbox, shown in the session. Verify that the command ran in the sandbox and not on the host. ``` ### Caveat The README gives the proxy and local operations manager pattern as an example, so the setup steps for a specific sandbox must be worked out from the source. Risk, remote script: Bash commands run inside the remote sandbox, so limit the sandbox's network and file access to what the task needs. Dual-use notice: only use this on software, sites and systems you own or are authorized to test. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/unreallabsai-unreal-agent/run-agent-tools-in-remote-sandbox/