William
William is not enabled by default
This is a limited access feature. Contact the nunu team to enable it for your project.
William is your helpful AI-Assistant that sits on top of nexus.
Rather than navigating nexus manually, you can describe what you want and William takes care of it - from writing a new test case to inspecting a live agent run.
William has full read/write access to your project’s entities including tests, test plans, shared steps, knowledge files, runs, configs, and builds.
Example Use Cases
- “Write a verification test that checks the onboarding flow for new users.”
- “Analyze the last test plan run and summarize why tests failed.”
- “Update all my smoke tests to use the new shared setup step.”
Capabilities
| Category | What William Can Do |
|---|---|
| Tests | Create, update, delete test cases across all three types (verification, discovery, task). Describe what you want to test and he’ll produce a properly structured test with goals, expected results, and sensible defaults. |
| Test Plans | Create and manage test plans — ordered collections of test cases you can run together. |
| Shared Steps | Create, update, and delete shared steps for reuse across tests. |
| Runs | Trigger runs, list past runs, inspect live runs in real time (including screenshots), and read artifacts like logs and output files. |
| Knowledge | Create, update, move, and delete knowledge files scoped to the project, an agent config, or a specific test case. |
| Configs | Look up agent configs and deployment configs (read-only). |
| Builds | List and reference available builds. |
| Search | Search across tests, shared steps, and past runs, and retrieve details about any entity in your project. |
| Docs | Look up documentation on any Nexus feature or nunu.ai platform topic. |
Limitations
William is focused on tasks within the Nexus platform. He cannot:
- Know about your internal information (Provide the necessary context!)
- Access external systems or URLs (unless documented in your knowledge files)
- Modify your game code or SDK integration
- Create or modify agent/deployment configs directly (read-only)
- Execute arbitrary code
Last updated on